[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Next Issue
Volume 13, June
Previous Issue
Volume 13, April
You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
7.1
2.8
Journals
Future Internet
Volume 13
Issue 5
share announcement

Future Internet, Volume 13, Issue 5 (May 2021) – 34 articles

Cover Story (view full-size image): This article details the design, development, and assessment of a wildlife monitoring application for IoT animal-repelling devices that is able to cover large areas, thanks to low-power wide area network (LPWAN), which bridges the gap between cellular technologies and short-range wireless technologies. This article first presents a detailed survey of LPWAN for smart agriculture and wildlife monitoring applications. We evaluate the performance of LoRa (the global de-facto LPWAN) transmission technology, operating in the 433 and 868 MHz bands, aimed at wildlife monitoring in a forest vegetation area. Findings from this study show that achievable performance can greatly vary between the 433 and 868 MHz, and caution is required when taking numbers at face value, as this can have implications for IoT applications. View this paper
  • Issues are regarded as officially published after their release is announced to the table of contents alert mailing list.
  • You may sign up for e-mail alerts to receive table of contents of newly released issues.
  • PDF is the official format for papers published in both, html and pdf forms. To view the papers in pdf format, click on the "PDF Full-text" link, and use the free Adobe Reader to open them.
Order results
Result details
Section
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
24 pages, 893 KiB  
Article
Generating Synthetic Training Data for Supervised De-Identification of Electronic Health Records
by Claudia Alessandra Libbi, Jan Trienes, Dolf Trieschnigg and Christin Seifert
Future Internet 2021, 13(5), 136; https://doi.org/10.3390/fi13050136 - 20 May 2021
Cited by 18 | Viewed by 5479
Abstract
A major hurdle in the development of natural language processing (NLP) methods for Electronic Health Records (EHRs) is the lack of large, annotated datasets. Privacy concerns prevent the distribution of EHRs, and the annotation of data is known to be costly and cumbersome. [...] Read more.
A major hurdle in the development of natural language processing (NLP) methods for Electronic Health Records (EHRs) is the lack of large, annotated datasets. Privacy concerns prevent the distribution of EHRs, and the annotation of data is known to be costly and cumbersome. Synthetic data presents a promising solution to the privacy concern, if synthetic data has comparable utility to real data and if it preserves the privacy of patients. However, the generation of synthetic text alone is not useful for NLP because of the lack of annotations. In this work, we propose the use of neural language models (LSTM and GPT-2) for generating artificial EHR text jointly with annotations for named-entity recognition. Our experiments show that artificial documents can be used to train a supervised named-entity recognition model for de-identification, which outperforms a state-of-the-art rule-based baseline. Moreover, we show that combining real data with synthetic data improves the recall of the method, without manual annotation effort. We conduct a user study to gain insights on the privacy of artificial text. We highlight privacy risks associated with language models to inform future research on privacy-preserving automated text generation and metrics for evaluating privacy-preservation during text generation. Full article
(This article belongs to the Special Issue Natural Language Engineering: Methods, Tasks and Applications)
►▼ Show Figures

Figure 1

Figure 1
<p>Illustrative example comparing standard text generation with the approach taken in this paper. We introduce special tokens to delimit protected health information (PHI). These tokens can be learned and generated like any other token by the language models. A prompt of three tokens defines the initial context.</p> Full article ">Figure 2
<p>Overview of this study. (1) Raw, EHR text is automatically de-identified and annotated with in-text PHI labels. (2) Pre-processed text is used to train two “structure-aware” language models: an LSTM and GPT-2. (3) Using different decoding strategies, two synthetic corpora are generated from each language model. (4) Synthetic text is evaluated regarding utility and privacy. (4.1) Utility is measured by comparing the performance of machine learning models trained on synthetic data with models trained on real data. (4.2) For the privacy evaluation, ROUGE n-gram overlap and retrieval-based BM25 scoring is used to select the most similar real documents. Afterwards, the synthetic-real document pairs are presented to participants in a user study.</p> Full article ">Figure 3
<p>Overview of extrinsic evaluation procedure. We compare three settings: (1) a model trained on real data (baseline), (2) a “mixed” case, where we extend real data with synthetic data, and (3) only synthetic training data. All models were tested on real data (gold annotations). This evaluation setup extends Ive et al. [<a href="#B15-futureinternet-13-00136" class="html-bibr">15</a>] by step (2).</p> Full article ">Figure 4
<p>Illustration of method used to compile a set of similar synthetic-real document pairs for the privacy evaluation. For each synthetic document, we retrieve the most similar source documents from the real data, based on ROUGE n-gram overlap and BM25. The set <math display="inline"><semantics> <mrow> <mi>S</mi> <mi>R</mi> </mrow> </semantics></math> contains the pooled result of this matching process, such that each synthetic document appears in three separate pairings: once with the top ROUGE-3 match, once with the top ROUGE-5 match and once with the top BM25 match.</p> Full article ">Figure 5
<p>How well do the synthetic corpora reflect the real PHI distribution? This figure shows the differences to the PHI distribution of the language model training data (cf. <a href="#futureinternet-13-00136-t001" class="html-table">Table 1</a>).</p> Full article ">Figure 6
<p>Text sample from the GPT-beam corpus (translated from Dutch, PHI highlighted and replaced with random identifiers). The structure of the generated text resembles a template that nurses used in the real data to report on dental hygiene of a patient.</p> Full article ">Figure 7
<p>Inter-participant agreement (count of answer given) for the user study on privacy.</p> Full article ">Figure A1
<p>PHI distribution of the synthetic corpora compared to the language modeling corpus.</p> Full article ">Figure A2
<p>Raw increase/decrease in relative PHI frequency per synthetic corpus compared with the PHI distribution of the language modeling corpus. Raw numbers for <a href="#futureinternet-13-00136-f005" class="html-fig">Figure 5</a>.</p> Full article ">Figure A3
<p>Entity-level F1-score for varying LSTM-p training set sizes. The full training set (100%) consists of all training and validation documents in LSTM-p. The F1-score is measured on the <span class="html-italic">NUT</span> test set. For each subset size, we train/test each model 3 times. The line shows the averaged scores along with the 95% confidence interval.</p> Full article ">Figure A4
<p>Annotation guidelines for the privacy user study.</p> Full article ">
17 pages, 6126 KiB  
Article
Cognitive Profiling of Nodes in 6G through Multiplex Social Network and Evolutionary Collective Dynamics
by Marialisa Scatá, Barbara Attanasio and Aurelio La Corte
Future Internet 2021, 13(5), 135; https://doi.org/10.3390/fi13050135 - 20 May 2021
Cited by 2 | Viewed by 2659
Abstract
Complex systems are fully described by the connectedness of their elements studying how these develop a collective behavior, interacting with each other following their inner features, and the structure and dynamics of the entire system. The forthcoming 6G will attempt to rewrite the [...] Read more.
Complex systems are fully described by the connectedness of their elements studying how these develop a collective behavior, interacting with each other following their inner features, and the structure and dynamics of the entire system. The forthcoming 6G will attempt to rewrite the communication networks’ perspective, focusing on a radical revolution in the way entities and technologies are conceived, integrated and used. This will lead to innovative approaches with the aim of providing new directions to deal with future network challenges posed by the upcoming 6G, thus the complex systems could become an enabling set of tools and methods to design a self-organized, resilient and cognitive network, suitable for many application fields, such as digital health or smart city living scenarios. Here, we propose a complex profiling approach of heterogeneous nodes belonging to the network with the goal of including the multiplex social network as a mathematical representation that enables us to consider multiple types of interactions, the collective dynamics of diffusion and competition, through social contagion and evolutionary game theory, and the mesoscale organization in communities to drive learning and cognition. Through a framework, we detail the step by step modeling approach and show and discuss our findings, applying it to a real dataset, by demonstrating how the proposed model allows us to detect deeply complex knowable roles of nodes. Full article
(This article belongs to the Section Techno-Social Smart Systems)
►▼ Show Figures

Figure 1

Figure 1
<p>Multiplex representation mined from 6G subsystems. This figure shows the schematic representation through the multiplex network of a 6G scenario, including different types of sub-networks, various interactions and the evolutionary dynamics of collective phenomena.</p> Full article ">Figure 2
<p>Modeling Approach: a technical scheme of nodes’ profiling in a 6G network scenario. The figure graphically describes the various steps and the key aspects to profile the elements that constitute the 6G scenario into a complex network heterogeneous environment of cognitive elements. The approach starts from the application of the complexity and the multiplex representation and analysis to extract structural features as detailed, respectively, in (<b>a</b>,<b>b</b>). The scheme shows the collective profiling based on the analysis of diffusion and competition, representing two co-evolving and interdependent dynamics, as, respectively, detailed in (<b>c</b>,<b>d</b>). The community profiling, as in (<b>e</b>,<b>f</b>), is characterized by the detection of communities based on node and link hierarchical clustering obtained through a multilink graph.</p> Full article ">Figure 3
<p>Multiplex Network Representation. The figure describes both layers of the multiplex network <span class="html-italic">M</span>. The layer <math display="inline"><semantics> <msub> <mi>l</mi> <mn>1</mn> </msub> </semantics></math> is obtained from a real contact case, as explained in <a href="#sec3dot2-futureinternet-13-00135" class="html-sec">Section 3.2</a>, while the layer <math display="inline"><semantics> <msub> <mi>l</mi> <mn>2</mn> </msub> </semantics></math>, follows a theoretical scheme. Here, <math display="inline"><semantics> <msub> <mi>l</mi> <mn>2</mn> </msub> </semantics></math> is represented as a scale-free network. We shed light on the strength of each node (as the nodes’ size), and the awareness of each node (as the nodes’ color).</p> Full article ">Figure 4
<p>Density of Involved Nodes <math display="inline"><semantics> <mrow> <mi mathvariant="bold-italic">i</mi> <mo>(</mo> <mi mathvariant="bold-italic">t</mi> <mo>)</mo> </mrow> </semantics></math> versus time <math display="inline"><semantics> <mi mathvariant="bold-italic">t</mi> </semantics></math>, on the two processes in co-evolution. The outbreak of spreading in time is subjected to statistical variability. After a common exponential growth, we shed light on how the density of involved (“involved” means informed or infected) nodes <math display="inline"><semantics> <mrow> <mi mathvariant="bold-italic">i</mi> <mo>(</mo> <mi mathvariant="bold-italic">t</mi> <mo>)</mo> </mrow> </semantics></math> change in two different cases (respectively, orange and red) of the co-evolving processes that can converge to zero (<math display="inline"><semantics> <mrow> <msub> <mi mathvariant="bold-italic">S</mi> <mi mathvariant="bold-italic">h</mi> </msub> <mi mathvariant="bold-italic">IR</mi> <mo>−</mo> <msub> <mi mathvariant="bold-italic">U</mi> <mi mathvariant="bold-italic">h</mi> </msub> <mi mathvariant="bold-italic">AF</mi> </mrow> </semantics></math>) or to a constant number (<math display="inline"><semantics> <mrow> <msub> <mi mathvariant="bold-italic">S</mi> <mi mathvariant="bold-italic">h</mi> </msub> <mi mathvariant="bold-italic">IS</mi> <mo>−</mo> <msub> <mi mathvariant="bold-italic">U</mi> <mi mathvariant="bold-italic">h</mi> </msub> <mi mathvariant="bold-italic">AU</mi> </mrow> </semantics></math>).</p> Full article ">Figure 5
<p>Roles of Nodes from Structural Analysis. We show four panels to shed light on the roles of each node in <span class="html-italic">M</span> as a result of the structural profiling. At the top, the two heat-maps show with a color-code the value of degree <math display="inline"><semantics> <msubsup> <mi>k</mi> <mrow> <mi>i</mi> </mrow> <mi>α</mi> </msubsup> </semantics></math>, strength <math display="inline"><semantics> <msubsup> <mi>s</mi> <mrow> <mi>i</mi> </mrow> <mi>α</mi> </msubsup> </semantics></math> and overlapping degree <math display="inline"><semantics> <msub> <mi>o</mi> <mi>i</mi> </msub> </semantics></math> for each node, from the largest (darkest) to the smallest (brightest) in both cases of scale-free (SF) and small-world networks (SW) as assumption for the graph of the second layer of the multiplex that describes a social-based interaction graph, jointly with the first layer based on a real contact network (RC). Below, the other two panels show two cartography obtained by plotting, for each node, the multiplex participation coefficient <math display="inline"><semantics> <msub> <mi>P</mi> <mi>i</mi> </msub> </semantics></math> versus the Z-score of the total overlapping degree <math display="inline"><semantics> <msub> <mi>o</mi> <mi>i</mi> </msub> </semantics></math> in the cases of a theoretical distribution with high standard deviation <math display="inline"><semantics> <mi>δ</mi> </semantics></math> for low homophily and low standard deviation <math display="inline"><semantics> <mi>δ</mi> </semantics></math> for high homophily, referring to the distribution of similarity features among nodes. In both plots, comparing the cases <math display="inline"><semantics> <mrow> <mi>R</mi> <mi>C</mi> <mo>−</mo> <mi>S</mi> <mi>F</mi> </mrow> </semantics></math> and <math display="inline"><semantics> <mrow> <mi>R</mi> <mi>C</mi> <mo>−</mo> <mi>S</mi> <mi>W</mi> </mrow> </semantics></math>, we identify specific roles of nodes put in evidence by the plots, in terms of focused, mixed, multiplex, regular or hubs nodes.</p> Full article ">Figure 6
<p>Roles of Nodes from Collective Dynamics Analysis. We show four plots to shed light on the roles of each node of the population considered for the multiplex network <span class="html-italic">M</span> as a result of the collective profiling. The plots display, in the plane <math display="inline"><semantics> <mi>λ</mi> </semantics></math>-<math display="inline"><semantics> <mi>β</mi> </semantics></math>, the trend of involved rate in function of the awareness rate, comparing both cases of scale-free (SF) and small-world (SW) networks as assumption for the graph of the second layer of the multiplex that describe a social-based interaction graph, jointly with the first layer based on a real contact network (RC), showing in red color the <math display="inline"><semantics> <mrow> <mi>R</mi> <mi>C</mi> <mo>−</mo> <mi>S</mi> <mi>F</mi> </mrow> </semantics></math> case, while in yellow the <math display="inline"><semantics> <mrow> <mi>R</mi> <mi>C</mi> <mo>−</mo> <mi>S</mi> <mi>W</mi> </mrow> </semantics></math> case. In accordance with the model and the interdependence between the evolutionary dynamics linked to the co-evolving spreading processes and the evolutionary game theory, we set and compare the four social dilemmas (PD-SD-HG-SH) displaying the trend of the spreading dynamics in relation to the number of cooperation <math display="inline"><semantics> <msub> <mi>n</mi> <mi>c</mi> </msub> </semantics></math> for each node (as dot size variation) of the multiplex network <span class="html-italic">M</span>.</p> Full article ">Figure 7
<p>Roles of Nodes from Community Analysis. In the (<b>a</b>,<b>b</b>) panel, we display the findings of the hierarchical clustering of nodes based on the score function <math display="inline"><semantics> <mrow> <mi>h</mi> <msub> <mrow> <msub> <mi>c</mi> <mrow> <mi>i</mi> <mi>j</mi> </mrow> </msub> </mrow> <mrow> <mrow> <mo>|</mo> </mrow> <mover accent="true"> <mi>G</mi> <mo>^</mo> </mover> </mrow> </msub> </mrow> </semantics></math>. In (<b>a</b>), we show the cluster dendrogram, with each colored square shedding lights on different node communities, determined by cutting the dendrogram at a specific height. In panel (<b>b</b>), we display the visualization of nodes communities using graph highlighting the six communities of the weighted aggregated graphs <math display="inline"><semantics> <mover accent="true"> <mi>G</mi> <mo>^</mo> </mover> </semantics></math>, whose interaction are referred to existing multilinks between a pair of nodes, and it is extracted from the multiplex network <span class="html-italic">M</span>. Moreover, we underline the multidegree values <math display="inline"><semantics> <msubsup> <mi>k</mi> <mi>i</mi> <mover accent="true"> <mi>m</mi> <mo>→</mo> </mover> </msubsup> </semantics></math> for each node in accordance with their size. In the panel (<b>c</b>,<b>d</b>), we exhibit the findings of a link-based clustering of the graph <math display="inline"><semantics> <mover accent="true"> <mi>G</mi> <mo>^</mo> </mover> </semantics></math> of the multiplex <span class="html-italic">M</span>, to detect the links communities. In (<b>c</b>), we show the cluster dendrogram highlighting the sixteen clusters detected. In (<b>d</b>), we illustrate the visualization of links communities using graph. Panel (<b>d</b>) shows a Spencer circle layout with numbers around the circumference of the circle that are referred to community ID.</p> Full article ">
22 pages, 1980 KiB  
Article
Comparative Study of Distributed Consensus Gossip Algorithms for Network Size Estimation in Multi-Agent Systems
by Martin Kenyeres and Jozef Kenyeres
Future Internet 2021, 13(5), 134; https://doi.org/10.3390/fi13050134 - 18 May 2021
Cited by 15 | Viewed by 3843
Abstract
Determining the network size is a critical process in numerous areas (e.g., computer science, logistic, epidemiology, social networking services, mathematical modeling, demography, etc.). However, many modern real-world systems are so extensive that measuring their size poses a serious challenge. Therefore, the algorithms for [...] Read more.
Determining the network size is a critical process in numerous areas (e.g., computer science, logistic, epidemiology, social networking services, mathematical modeling, demography, etc.). However, many modern real-world systems are so extensive that measuring their size poses a serious challenge. Therefore, the algorithms for determining/estimating this parameter in an effective manner have been gaining popularity over the past decades. In the paper, we analyze five frequently applied distributed consensus gossip-based algorithms for network size estimation in multi-agent systems (namely, the Randomized gossip algorithm, the Geographic gossip algorithm, the Broadcast gossip algorithm, the Push-Sum protocol, and the Push-Pull protocol). We examine the performance of the mentioned algorithms with bounded execution over random geometric graphs by applying two metrics: the number of sent messages required for consensus achievement and the estimation precision quantified as the median deviation from the real value of the network size. The experimental part consists of two scenarios—the consensus achievement is conditioned by either the values of the inner states or the network size estimates—and, in both scenarios, either the best-connected or the worst-connected agent is chosen as the leader. The goal of this paper is to identify whether all the examined algorithms are applicable to estimating the network size, which algorithm provides the best performance, how the leader selection can affect the performance of the algorithms, and how to most effectively configure the applied stopping criterion. Full article
(This article belongs to the Special Issue Modern Trends in Multi-Agent Systems)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>General structure of agent forming multi-agent system.</p> Full article ">Figure 2
<p>Various classifications of data aggregation methods.</p> Full article ">Figure 3
<p>Example of distributed architecture.</p> Full article ">Figure 4
<p>Example of random geometric graphs used in our experiments.</p> Full article ">Figure 5
<p>Comparison of inner states/arithmetic mean estimates before and after consensus is achieved in multi-agent system.</p> Full article ">Figure 6
<p>Comparison of gossip Push-Sum protocol with deterministic Max-Degree weights algorithm over three runs—evolution of arithmetic mean estimates.</p> Full article ">Figure 7
<p>Comparison of gossip Push-Sum protocol with deterministic Max-Degree weights algorithm—distribution of convergence rate over 100,000 runs.</p> Full article ">Figure 8
<p>Comparison of evolution of arithmetic mean estimates with evolution of network size estimates—pairwise gossip algorithm is applied.</p> Full article ">Figure 9
<p>Comparison of final network size estimates in both examined scenarios—pairwise gossip algorithm is applied.</p> Full article ">Figure 10
<p>Comparison of the number of sent messages required for consensus achievement of all algorithms in both scenarios for various configurations of implemented stopping criterion—best-connected agent is leader.</p> Full article ">Figure 11
<p>Comparison of the estimation precision of all algorithms in both scenarios for various configurations of implemented stopping criterion—best-connected agent is leader.</p> Full article ">Figure 12
<p>Comparison of the number of sent messages required for consensus achievement of all algorithms in both scenarios for various configurations of implemented stopping criterion—worst-connected agent is leader.</p> Full article ">Figure 13
<p>Comparison of the estimation precision of all algorithms in both scenarios for various configurations of implemented stopping criterion—worst-connected agent is leader.</p> Full article ">Figure 14
<p>Distribution of number of sent messages in both scenarios for various configurations of implemented stopping criterion and with the best-connected agent as leader.</p> Full article ">
14 pages, 10955 KiB  
Article
Blockchain Implementation Method for Interoperability between CBDCs
by Hyunjun Jung and Dongwon Jeong
Future Internet 2021, 13(5), 133; https://doi.org/10.3390/fi13050133 - 18 May 2021
Cited by 21 | Viewed by 5826
Abstract
Central Bank Digital Currency (CBDC) is a digital currency issued by a central bank. Motivated by the financial crisis and prospect of a cashless society, countries are researching CBDC. Recently, global consideration has been given to paying basic income to avoid consumer sentiment [...] Read more.
Central Bank Digital Currency (CBDC) is a digital currency issued by a central bank. Motivated by the financial crisis and prospect of a cashless society, countries are researching CBDC. Recently, global consideration has been given to paying basic income to avoid consumer sentiment shrinkage and recession due to epidemics. CBDC is coming into the spotlight as the way to manage the public finance policy of nations comprehensively. CBDC is studied by many countries. The bank of the Bahamas released Sand Dollar. Each country’s central bank should consider the situation in which CBDCs are exchanged. The transaction of the CDDB is open data. Transaction registers CBDC exchange information of the central bank in the blockchain. Open data on currency exchange between countries will provide information on the flow of money between countries. This paper proposes a blockchain system and management method based on the ISO/IEC 11179 metadata registry for exchange between CBDCs that records transactions between registered CBDCs. Each country’s CBDC will have a different implementation and time of publication. We implement the blockchain system and experiment with the operation method, measuring the block generation time of blockchains using the proposed method. Full article
(This article belongs to the Special Issue Open Data and Artificial Intelligence)
►▼ Show Figures

Figure 1

Figure 1
<p>ISO/IEC 11179 Metadata Registry—Data Element.</p> Full article ">Figure 2
<p>Data Element Structure of MDR.</p> Full article ">Figure 3
<p>Data Element Concept example of MDR.</p> Full article ">Figure 4
<p>Conceptual Domain and Value domain example of MDR.</p> Full article ">Figure 5
<p>Registration procedure of CBDC.</p> Full article ">Figure 6
<p>MDR-based CBDC blockchain model.</p> Full article ">Figure 7
<p>Registration procedure of CBDC.</p> Full article ">Figure 8
<p>Create account and block in bitcoin.</p> Full article ">Figure 9
<p>Convert input data to Hex.</p> Full article ">Figure 10
<p>Create raw transaction in bitcoin.</p> Full article ">Figure 11
<p>Sign and Send Raw Transaction in bitcoin.</p> Full article ">Figure 12
<p>Insert transaction to block in bitcoin.</p> Full article ">Figure 13
<p>Decode raw transaction in bitcoin.</p> Full article ">Figure 14
<p>Transaction generation in Ethereum.</p> Full article ">Figure 15
<p>List of transactions stored in the transaction pool in Ethereum.</p> Full article ">Figure 16
<p>Generated block contents in Ethereum.</p> Full article ">Figure 17
<p>Generated transaction contents in Ethereum.</p> Full article ">Figure 18
<p>Block generation time in blockchains ((<b>A</b>) Bitcoin; (<b>B</b>) Ethereum).</p> Full article ">
20 pages, 522 KiB  
Review
Trust, but Verify: Informed Consent, AI Technologies, and Public Health Emergencies
by Brian Pickering
Future Internet 2021, 13(5), 132; https://doi.org/10.3390/fi13050132 - 18 May 2021
Cited by 21 | Viewed by 10429
Abstract
To use technology or engage with research or medical treatment typically requires user consent: agreeing to terms of use with technology or services, or providing informed consent for research participation, for clinical trials and medical intervention, or as one legal basis for processing [...] Read more.
To use technology or engage with research or medical treatment typically requires user consent: agreeing to terms of use with technology or services, or providing informed consent for research participation, for clinical trials and medical intervention, or as one legal basis for processing personal data. Introducing AI technologies, where explainability and trustworthiness are focus items for both government guidelines and responsible technologists, imposes additional challenges. Understanding enough of the technology to be able to make an informed decision, or consent, is essential but involves an acceptance of uncertain outcomes. Further, the contribution of AI-enabled technologies not least during the COVID-19 pandemic raises ethical concerns about the governance associated with their development and deployment. Using three typical scenarios—contact tracing, big data analytics and research during public emergencies—this paper explores a trust-based alternative to consent. Unlike existing consent-based mechanisms, this approach sees consent as a typical behavioural response to perceived contextual characteristics. Decisions to engage derive from the assumption that all relevant stakeholders including research participants will negotiate on an ongoing basis. Accepting dynamic negotiation between the main stakeholders as proposed here introduces a specifically socio–psychological perspective into the debate about human responses to artificial intelligence. This trust-based consent process leads to a set of recommendations for the ethical use of advanced technologies as well as for the ethical review of applied research projects. Full article
(This article belongs to the Special Issue Human–Computer Interaction Models and Experiences for Internet of Things Systems and Edge Computing)
►▼ Show Figures

Figure 1

Figure 1
<p>Schematic representation of the Theory of Planned Behavior [<a href="#B18-futureinternet-13-00132" class="html-bibr">18</a>].</p> Full article ">Figure 2
<p>TPB Interpretation of Behaviours associated with (<b>a</b>) <span class="html-italic">Terms of Use</span> and (<b>b</b>) <span class="html-italic">Research Consent</span>.</p> Full article ">Figure 3
<p>A Schematic Representation of Trust Behaviours.</p> Full article ">
19 pages, 1103 KiB  
Article
A Hierarchical Cache Size Allocation Scheme Based on Content Dissemination in Information-Centric Networks
by Hongyu Liu and Rui Han
Future Internet 2021, 13(5), 131; https://doi.org/10.3390/fi13050131 - 15 May 2021
Cited by 4 | Viewed by 2491
Abstract
With the rapid growth of mass content retrieval on the Internet, Information-Centric Network (ICN) has become one of the hotspots in the field of future network architectures. The in-network cache is an important feature of ICN. For better network performance in ICN, the [...] Read more.
With the rapid growth of mass content retrieval on the Internet, Information-Centric Network (ICN) has become one of the hotspots in the field of future network architectures. The in-network cache is an important feature of ICN. For better network performance in ICN, the cache size on each node should be allocated in proportion to its importance. However, in some current studies, the importance of cache nodes is usually determined by their location in the network topology, ignoring their roles in the actual content transmission process. In this paper, we focus on the allocation of cache size for each node within a given total cache space budget. We explore the impact of heterogeneous cache allocation on content dissemination under the same ICN infrastructure and we quantify the importance of nodes from content dissemination and network topology. To this purpose, we implement a hierarchy partitioning method based on content dissemination, then we formulate a set of weight calculation methods for these hierarchies and to provide a per-node cache space allocation to allocate the total cache space budget to each node in the network. The performance of the scheme is evaluated on the Garr topology, and the average hit ratio, latency, and load are compared to show that the proposed scheme has better performance in these aspects than other schemes. Full article
(This article belongs to the Section Network Virtualization and Edge/Fog Computing)
►▼ Show Figures

Figure 1

Figure 1
<p>Present Research on Cache in ICN.</p> Full article ">Figure 2
<p>The Arbitrary Graph Structure Network Model.</p> Full article ">Figure 3
<p>Comparison of Two Caching Schemes.</p> Full article ">Figure 4
<p>Betweenness Centrality of Network Nodes.</p> Full article ">Figure 5
<p>The Content Domains.</p> Full article ">Figure 6
<p>The Data Paths in Content Domain A.</p> Full article ">Figure 7
<p>Performance of Cache Hit Ratio.</p> Full article ">Figure 8
<p>Performance of Request Latency.</p> Full article ">Figure 9
<p>Performance of Content Server Load.</p> Full article ">
28 pages, 1312 KiB  
Article
A Digital Currency Architecture for Privacy and Owner-Custodianship
by Geoffrey Goodell, Hazem Danny Al-Nakib and Paolo Tasca
Future Internet 2021, 13(5), 130; https://doi.org/10.3390/fi13050130 - 14 May 2021
Cited by 14 | Viewed by 6283
Abstract
In recent years, electronic retail payment mechanisms, especially e-commerce and card payments at the point of sale, have increasingly replaced cash in many developed countries. As a result, societies are losing a critical public retail payment option, and retail consumers are losing important [...] Read more.
In recent years, electronic retail payment mechanisms, especially e-commerce and card payments at the point of sale, have increasingly replaced cash in many developed countries. As a result, societies are losing a critical public retail payment option, and retail consumers are losing important rights associated with using cash. To address this concern, we propose an approach to digital currency that would allow people without banking relationships to transact electronically and privately, including both e-commerce purchases and point-of-sale purchases that are required to be cashless. Our proposal introduces a government-backed, privately-operated digital currency infrastructure to ensure that every transaction is registered by a bank or money services business, and it relies upon non-custodial wallets backed by privacy-enhancing technology, such as blind signatures or zero-knowledge proofs, to ensure that transaction counterparties are not revealed. Our approach to digital currency can also facilitate more efficient and transparent clearing, settlement, and management of systemic risk. We argue that our system can restore and preserve the salient features of cash, including privacy, owner-custodianship, fungibility, and accessibility, while also preserving fractional reserve banking and the existing two-tiered banking system. We also show that it is possible to introduce regulation of digital currency transactions involving non-custodial wallets that unconditionally protect the privacy of end-users. Full article
(This article belongs to the Special Issue Blockchain Security and Privacy)
►▼ Show Figures

Figure 1

Figure 1
<p>Taxonomy of digital money systems.</p> Full article ">Figure 2
<p>Schematic representation of the system design.</p> Full article ">Figure 3
<p>Schematic representation of a typical user engagement lifecycle.</p> Full article ">Figure 4
<p>Schematic representation of a user depositing CBDC into a bank account.</p> Full article ">Figure 5
<p>Schematic representation of a mediated transaction between consumers.</p> Full article ">Figure 6
<p>Schematic representation of a disbursement to a retail user with a non-custodial wallet.</p> Full article ">
19 pages, 418 KiB  
Review
Hashtag Recommendation Methods for Twitter and Sina Weibo: A Review
by Areej Alsini, Du Q. Huynh and Amitava Datta
Future Internet 2021, 13(5), 129; https://doi.org/10.3390/fi13050129 - 14 May 2021
Cited by 11 | Viewed by 5154
Abstract
Hashtag recommendation suggests hashtags to users while they write microblogs in social media platforms. Although researchers have investigated various methods and factors that affect the performance of hashtag recommendations in Twitter and Sina Weibo, a systematic review of these methods is lacking. The [...] Read more.
Hashtag recommendation suggests hashtags to users while they write microblogs in social media platforms. Although researchers have investigated various methods and factors that affect the performance of hashtag recommendations in Twitter and Sina Weibo, a systematic review of these methods is lacking. The objectives of this study are to present a comprehensive overview of research on hashtag recommendation for tweets and present insights from previous research papers. In this paper, we search for articles related to our research between 2010 and 2020 from CiteSeer, IEEE Xplore, Springer and ACM digital libraries. From the 61 articles included in this study, we notice that most of the research papers were focused on the textual content of tweets instead of other data. Furthermore, collaborative filtering methods are seldom used solely in hashtag recommendation. Taking this perspective, we present a taxonomy of hashtag recommendation based on the research methodologies that have been used. We provide a critical review of each of the classes in the taxonomy. We also discuss the challenges remaining in the field and outline future research directions in this area of study. Full article
(This article belongs to the Special Issue Social Networks Analysis and Mining)
►▼ Show Figures

Graphical abstract

Graphical abstract
Full article ">Figure 1
<p>Flowchart of the steps used in the literature search.</p> Full article ">Figure 2
<p>Distribution of papers from 2010 to 2020.</p> Full article ">Figure 3
<p>Distribution of the eligible papers on hashtag recommendation from 2010 to 2020 with respect to: (<b>a</b>) the three main categories of methods under our taxonomy; (<b>b</b>) the category of text-based methods; and (<b>c</b>) the category of hybrid user-based methods.</p> Full article ">Figure 4
<p>Our proposed taxonomy of the hashtag recommendation methods.</p> Full article ">
19 pages, 7604 KiB  
Article
Research on Task-Oriented Computation Offloading Decision in Space-Air-Ground Integrated Network
by Jun Liu, Xiaohui Lian and Chang Liu
Future Internet 2021, 13(5), 128; https://doi.org/10.3390/fi13050128 - 13 May 2021
Cited by 4 | Viewed by 2642
Abstract
In Space–Air–Ground Integrated Networks (SAGIN), computation offloading technology is a new way to improve the processing efficiency of node tasks and improve the limitation of computing storage resources. To solve the problem of large delay and energy consumption cost of task computation offloading, [...] Read more.
In Space–Air–Ground Integrated Networks (SAGIN), computation offloading technology is a new way to improve the processing efficiency of node tasks and improve the limitation of computing storage resources. To solve the problem of large delay and energy consumption cost of task computation offloading, which caused by the complex and variable network offloading environment and a large amount of offloading tasks, a computation offloading decision scheme based on Markov and Deep Q Networks (DQN) is proposed. First, we select the optimal offloading network based on the characteristics of the movement of the task offloading process in the network. Then, the task offloading process is transformed into a Markov state transition process to build a model of the computational offloading decision process. Finally, the delay and energy consumption weights are introduced into the DQN algorithm to update the computation offloading decision process, and the optimal offloading decision under the low cost is achieved according to the task attributes. The simulation results show that compared with the traditional Lyapunov-based offloading decision scheme and the classical Q-learning algorithm, the delay and energy consumption are respectively reduced by 68.33% and 11.21%, under equal weights when the offloading task volume exceeds 500 Mbit. Moreover, compared with offloading to edge nodes or backbone nodes of the network alone, the proposed mixed offloading model can satisfy more than 100 task requests with low energy consumption and low delay. It can be seen that the computation offloading decision proposed in this paper can effectively reduce the delay and energy consumption during the task computation offloading in the Space–Air–Ground Integrated Network environment, and can select the optimal offloading sites to execute the tasks according to the characteristics of the task itself. Full article
(This article belongs to the Section Big Data and Augmented Intelligence)
►▼ Show Figures

Figure 1

Figure 1
<p>Architecture of the Space–Earth Integration Network. The structure of the Space–Air–Ground Integrated Network is given from the perspective of network nodes, and the space-based backbone network is the core infrastructure of the space–ground integrated information network to meet the needs of various types of communications through the integration of space and ground networks.</p> Full article ">Figure 2
<p>Markov State Transition Model. The Space–Air–Ground Integrated Network is abstracted into an offloading network model represented by Markov states, with <span class="html-italic">i</span> denoting the task offloading location and circles de-noting the Markov states in each decision period. Taking the edge access network in the middle layer as an example, the whole Markovian decision process is given, which is represented by gray circles.</p> Full article ">Figure 3
<p>Network Switching Flowchart. This figure represents the network selection and switching process during task offloading. Network switching is always present during the whole task offloading process, always keeping the network where the current task is offloaded as the optimal network, so that the delay and energy cost of task offloading are minimized.</p> Full article ">Figure 4
<p>Task Offloading Process Movement Model. The connection boundary of the wireless network is a circular shape. Its center, O, is the transmitting and receiving antenna of the signal of the offload terminal device, which represents the terminal location of the current task offload of the network. The radius of the boundary circle (indicated by the dotted line) is the connection range of the infinite network, and the solid line indicates the moving path of the terminal equipment within the circle, where Point A represents the initial position of the offloading site when the task is unloaded, and Points B, C and D are the positions of the offloading site in the process of moving.</p> Full article ">Figure 5
<p>Comparison of total cost of delay and energy consumption weight at 0.5. Black indicates the computation offloading method based on Lyapunov algorithm, red indicates the computation offloading method based on classical Q-learning algorithm, and green indicates the computation offloading method of DQN algorithm proposed in this paper, comparing the cost of the three approaches when the delay and energy consumption are equally important.</p> Full article ">Figure 6
<p>Comparison of total cost when the delay weight is 0.9. Black indicates the computation offloading method based on Lyapunov algorithm, red indicates the computation offloading method based on classical Q-learning algorithm, and green indicates the computation offloading method of DQN algorithm proposed in this paper, comparing the time delay and energy consumption cost of the three approaches under experimentally sensitive task requirements.</p> Full article ">Figure 7
<p>The change in the number of subtasks switching frequency changes in offloading. Blue indicates the computation offloading based on Lyapunov algorithm, black indicates the computation offloading based on classical Q-learning algorithm, and red indicates the computation offloading method of DQN algorithm proposed in this paper. Compare the number of network switching with task amount during task offloading of the three approaches.</p> Full article ">Figure 8
<p>Comparison of energy consumption costs of different offloading sites. Blue indicates that all tasks are offloaded to the backbone nodes of the network, yellow indicates that all tasks are offloaded to the edge nodes of the network, and red indicates the mixed offloading mode proposed in this paper, where some tasks are offloaded to the backbone nodes and the others are offloaded to the edge nodes. Compare the energy consumption cost of the mixed computation offloading under a certain work request volume with that of the mode in which the particular offloading sites are selected separately.</p> Full article ">Figure 9
<p>Comparison of delay costs of different offloading sites. Blue indicates that all tasks are offloaded to the backbone nodes of the network, yellow indicates that all tasks are offloaded to the edge nodes of the network, and red indicates the mixed offloading mode proposed in this paper, where some tasks are offloaded to the backbone nodes and the others are offloaded to the edge nodes. Compare the delay cost of the mixed computation offloading under a certain work request volume with that of the mode in which a particular offloading sites are selected separately.</p> Full article ">
19 pages, 3738 KiB  
Article
User Acceptance of Smart Watch for Medical Purposes: An Empirical Study
by Rana Saeed Al-Maroof, Khadija Alhumaid, Ahmad Qasim Alhamad, Ahmad Aburayya and Said Salloum
Future Internet 2021, 13(5), 127; https://doi.org/10.3390/fi13050127 - 12 May 2021
Cited by 34 | Viewed by 9224
Abstract
This study aims to investigate the most effective and interesting variables that urge use of the smartwatch (SW) in a medical environment. To achieve this aim, the study was framed using an innovative and integrated research model, which is based on combining constructs [...] Read more.
This study aims to investigate the most effective and interesting variables that urge use of the smartwatch (SW) in a medical environment. To achieve this aim, the study was framed using an innovative and integrated research model, which is based on combining constructs from a well-established theoretical model’s TAM and other features that are critical to the effectiveness of SW which are content richness and personal innovativeness. The Technology Acceptance Model (TAM) is used to detect the determinants affecting the adoption of SW. The current study depends on an online questionnaire that is composed of (20) items. The questionnaire is distributed among a group of doctors, nurses, and administration staff in medical centers within the UAE. The total number of respondents is (325). The collected data were implemented to test the study model and the proposed constructs and hypotheses depending on the Smart PLS Software. The results of the current study show that the main constructs in the model contribute differently to the acceptance of SW. Based on the previous assumption, content richness and innovativeness are critical factors that enrich the user’s perceived usefulness. In addition, perceived ease of use was significantly predictive of either perceived usefulness or behavioral intention. Overall findings suggest that SW is in high demand in the medical field and is used as a common channel among doctors and their patients and it facilitates the role of transmitting information among its users. The outcomes of the current study indicate the importance of certain external factors for the acceptance of the technology. The genuine value of this study lies in the fact that it is based on a conceptual framework that emphasizes the close relationship between the TAM constructs of perceived usefulness and perceived ease of use to the construct of content richness, and innovativeness. Finally, this study helps us recognize the embedded motives for using SW in a medical environment, where the main motive is to enhance and facilitate the effective roles of doctors and patients. Full article
(This article belongs to the Special Issue The Future Internet of Medical Things)
►▼ Show Figures

Figure 1

Figure 1
<p>Research Model.</p> Full article ">Figure 2
<p>Path coefficient of the model (significant at ** <span class="html-italic">p</span> ≤ 0.01, * <span class="html-italic">p</span> &lt; 0.05).</p> Full article ">
15 pages, 1974 KiB  
Article
Reducing Videoconferencing Fatigue through Facial Emotion Recognition
by Jannik Rößler, Jiachen Sun and Peter Gloor
Future Internet 2021, 13(5), 126; https://doi.org/10.3390/fi13050126 - 12 May 2021
Cited by 23 | Viewed by 4489
Abstract
In the last 14 months, COVID-19 made face-to-face meetings impossible and this has led to rapid growth in videoconferencing. As highly social creatures, humans strive for direct interpersonal interaction, which means that in most of these video meetings the webcam is switched on [...] Read more.
In the last 14 months, COVID-19 made face-to-face meetings impossible and this has led to rapid growth in videoconferencing. As highly social creatures, humans strive for direct interpersonal interaction, which means that in most of these video meetings the webcam is switched on and people are “looking each other in the eyes”. However, it is far from clear what the psychological consequences of this shift to virtual face-to-face communication are and if there are methods to alleviate “videoconferencing fatigue”. We have studied the influence of emotions of meeting participants on the perceived outcome of video meetings. Our experimental setting consisted of 35 participants collaborating in eight teams over Zoom in a one semester course on Collaborative Innovation Networks in bi-weekly video meetings, where each team presented its progress. Emotion was tracked through Zoom face video snapshots using facial emotion recognition that recognized six emotions (happy, sad, fear, anger, neutral, and surprise). Our dependent variable was a score given after each presentation by all participants except the presenter. We found that the happier the speaker is, the happier and less neutral the audience is. More importantly, we found that the presentations that triggered wide swings in “fear” and “joy” among the participants are correlated with a higher rating. Our findings provide valuable input for online video presenters on how to conduct better and less tiring meetings; this will lead to a decrease in “videoconferencing fatigue”. Full article
(This article belongs to the Special Issue Social Networks Analysis and Mining)
►▼ Show Figures

Figure 1

Figure 1
<p>Overview of the data pre-processing steps. Each recording is divided into eight videos and each contains a different presentation. We then extracted one image per second for each presentation video. Finally, we extracted all faces from a single image and classified them into presenter, supervisor, and audience.</p> Full article ">Figure 2
<p>VGGs16 confusion matrix on the test set.</p> Full article ">Figure 3
<p>VGGs19 confusion matrix on the test set.</p> Full article ">Figure 4
<p>VGGs13 confusion matrix on the test set.</p> Full article ">Figure 5
<p>Xception confusion matrix on the test set.</p> Full article ">Figure 6
<p>(<b>a</b>) Illustration of the correlation between presentation score and the audience’s deviation in happiness. We can see that presentations which caused wide fluctuation in happiness among the spectators achieved higher scores (<b>b</b>,<b>c</b>) than presentations which did not cause fluctuation in happiness among the spectators (<b>d</b>,<b>e</b>).</p> Full article ">
11 pages, 1171 KiB  
Article
Towards Practical Applications in Modeling Blockchain System
by Andrei Vladyko, Anastasia Spirkina and Vasiliy Elagin
Future Internet 2021, 13(5), 125; https://doi.org/10.3390/fi13050125 - 12 May 2021
Cited by 10 | Viewed by 3016
Abstract
Like multiservice networks, blockchain technology is currently experiencing significant development because of its decentralization and ability to organize secure, seamless, reliable data exchange and storage. Due to the significant demand for the technology, there is a need to analyze the impact of these [...] Read more.
Like multiservice networks, blockchain technology is currently experiencing significant development because of its decentralization and ability to organize secure, seamless, reliable data exchange and storage. Due to the significant demand for the technology, there is a need to analyze the impact of these technology processes on network characteristics to predict traffic behavior and ensure required quality indicators, as well as on the stability of public communication network elements when blockchain technology operates. Conducting a full-scale experiment is a time-consuming task that cannot always be accomplished, so in this paper, the authors propose considering approaches to modeling these systems and, as an example, propose to use a simulation system to assess the performance of the network and its elements. Full article
(This article belongs to the Special Issue The Next Blockchain Wave Current Challenges and Future Prospects)
►▼ Show Figures

Figure 1

Figure 1
<p>Blockchain market growth dynamics.</p> Full article ">Figure 2
<p>Simulation of a system with blockchain nodes.</p> Full article ">Figure 3
<p>Simulation results of the blockchain node system: (<b>a</b>) Dependence of the proportion of packets on the waiting time for processing at the access node; (<b>b</b>) dependence of the proportion of packets on the delay at the access node; (<b>c</b>) dependence of the proportion of packets on the waiting time for processing at the aggregation node; (<b>d</b>) dependence of the proportion of packets on the delay at the aggregation node</p> Full article ">Figure 4
<p>Dependence of the proportion of packets at the end-to-end network delay of the system.</p> Full article ">
9 pages, 1410 KiB  
Article
Development of Knowledge Graph for Data Management Related to Flooding Disasters Using Open Data
by Jiseong Son, Chul-Su Lim, Hyoung-Seop Shim and Ji-Sun Kang
Future Internet 2021, 13(5), 124; https://doi.org/10.3390/fi13050124 - 11 May 2021
Cited by 13 | Viewed by 3723
Abstract
Despite the development of various technologies and systems using artificial intelligence (AI) to solve problems related to disasters, difficult challenges are still being encountered. Data are the foundation to solving diverse disaster problems using AI, big data analysis, and so on. Therefore, we [...] Read more.
Despite the development of various technologies and systems using artificial intelligence (AI) to solve problems related to disasters, difficult challenges are still being encountered. Data are the foundation to solving diverse disaster problems using AI, big data analysis, and so on. Therefore, we must focus on these various data. Disaster data depend on the domain by disaster type and include heterogeneous data and lack interoperability. In particular, in the case of open data related to disasters, there are several issues, where the source and format of data are different because various data are collected by different organizations. Moreover, the vocabularies used for each domain are inconsistent. This study proposes a knowledge graph to resolve the heterogeneity among various disaster data and provide interoperability among domains. Among disaster domains, we describe the knowledge graph for flooding disasters using Korean open datasets and cross-domain knowledge graphs. Furthermore, the proposed knowledge graph is used to assist, solve, and manage disaster problems. Full article
(This article belongs to the Special Issue Open Data and Artificial Intelligence)
►▼ Show Figures

Figure 1

Figure 1
<p>Proposed model.</p> Full article ">Figure 2
<p>Hierarchical skeleton structure between the Flooding concept and the existing knowledge graph and schema.</p> Full article ">Figure 3
<p>Principal knowledge graph with Flooding based on the open data.</p> Full article ">
16 pages, 3231 KiB  
Article
A Collaborative Merging Strategy with Lane Changing in Multilane Freeway On-Ramp Area with V2X Network
by Yanjun Shi, Hao Yu, Yijia Guo and Zhiheng Yuan
Future Internet 2021, 13(5), 123; https://doi.org/10.3390/fi13050123 - 10 May 2021
Cited by 7 | Viewed by 3677
Abstract
The merging area of the freeway is an area with a high incidence of traffic accidents. With the development of connected and automated vehicles (CAVs) and V2X technology, the traffic efficiency of freeway ramp areas has been significantly improved. However, current research mostly [...] Read more.
The merging area of the freeway is an area with a high incidence of traffic accidents. With the development of connected and automated vehicles (CAVs) and V2X technology, the traffic efficiency of freeway ramp areas has been significantly improved. However, current research mostly focuses on merging a single mainline lane and ramp, and there are few cases of multiple lanes. In this paper, we present a collaborative merging model with a rule-based lane-changing strategy in a V2X environment. First, the vehicle selects the appropriate gap to change lanes safely without affecting other vehicles. Meanwhile, we established a linear time discrete model to optimize the trajectory of vehicles in real-time. Finally, the proposed model and strategy were implemented in SUMO and Python. The simulation results showed that the merging model we proposed based on the lane-changing strategy had good performance in terms of the number of stops, average delay, and average speed. Full article
►▼ Show Figures

Figure 1

Figure 1
<p>Schematic illustration of a typical two-lane freeway merging area.</p> Full article ">Figure 2
<p>Regional division of ramp area on a two-lane freeway.</p> Full article ">Figure 3
<p>Schematic diagram of the position of target vehicle at time <span class="html-italic">t.</span></p> Full article ">Figure 4
<p>Schematic diagram of the target gap for lane changing in the lane-changing zone.</p> Full article ">Figure 5
<p>Vehicle conditions in the Outer Lane within the target gap for lane changing: (<b>a</b>) only one vehicle exists; (<b>b</b>) there are multiple vehicles.</p> Full article ">Figure 6
<p>Flow chart of vehicle lane changing in the lane-changing zone.</p> Full article ">Figure 7
<p>Simplified diagram of the trajectory optimization zone.</p> Full article ">Figure 8
<p>The first optimization in the trajectory optimization zone.</p> Full article ">Figure 9
<p>Schematic diagram of vehicle trajectory. The left side is the scenario without lane changing, and the right side is the lane-changing scenario.</p> Full article ">Figure 10
<p>Comparison of different cases under different demands and safety time gaps: (<b>a</b>) The number of stops, (<b>b</b>) average delay, and (<b>c</b>) average speed.</p> Full article ">
11 pages, 5367 KiB  
Article
An Income Model Using Historical Data, Power-Law Distributions and Monte Carlo Method for University Technology Transfer Offices
by Ken Polasko, Pedro Ponce and Arturo Molina
Future Internet 2021, 13(5), 122; https://doi.org/10.3390/fi13050122 - 6 May 2021
Viewed by 2563
Abstract
Engineering education pushes the creation of new technology to solve community problems. The process of technology transfer promotes educational innovation in universities, a vital process that can improve citizens’ quality of life in cities and rural communities. As a result, university technology transfer [...] Read more.
Engineering education pushes the creation of new technology to solve community problems. The process of technology transfer promotes educational innovation in universities, a vital process that can improve citizens’ quality of life in cities and rural communities. As a result, university technology transfer offices (TTOs) have to create strategies that motivate students and researchers to generate technology. Thus, a primary challenge that TTOs face is to know and communicate the income potential compared to their much more predictable and limited expense budgets. Institutional budgeting for a TTO’s growth would be simplified if the office were on a solid financial footing, i.e., breaking even or making a financial return. Many offices assume that income is unpredictable, that it is a lottery, luck, and more stakes in the fire improve the odds of hitting a winner, etc. These common assumptions or beliefs provide only a vague insight into how to move an intellectual property (IP) portfolio strategy forward. How can a TTO be assessed for quantitative value and not just be a cost center adding qualitative value? This paper illustrates the first steps to understanding how to project potential income versus a much more predictable expense budget, which would allow universities to improve their technology transfer strategy and results. As a result, TTOs would operate under a more sustainable IP portfolio strategy, promote educational innovation in universities, and generate a more significant community impact. Full article
(This article belongs to the Special Issue Education 4.0 in the Transformation of Universities: Educational and Research Applications)
►▼ Show Figures

Figure 1

Figure 1
<p>Typical distribution of venture capital returns.</p> Full article ">Figure 2
<p>The fiscal year 2016 Association of University Technology Managers (AUTM) data.</p> Full article ">Figure 3
<p>AUTM binned data plotted on a log–log scale.</p> Full article ">Figure 4
<p>Two very different US research universities, yet, the two schools’ royalty incomes are similar and can be modeled as power-law distributions. (<b>a</b>) Event number vs. royalty income; (<b>b</b>) income vs. frequency.</p> Full article ">Figure 5
<p>Make up different return categories for venture and two universities.</p> Full article ">Figure 6
<p>Monte Carlo simulation of 400 portfolio-driven events.</p> Full article ">Figure 7
<p>Portfolio return probability after 10,000 Monte Carlo simulations.</p> Full article ">Figure 8
<p>Impact of a lower quality portfolio: fewer power-law returns: 3.7 &gt; 1.0%.</p> Full article ">Figure 9
<p>The challenge of breaking even when 20% of the patented portfolio is licensed, and the cost burden of the 80% unlicensed portfolio is included.</p> Full article ">
23 pages, 713 KiB  
Article
Exploiting Machine Learning for Improving In-Memory Execution of Data-Intensive Workflows on Parallel Machines
by Riccardo Cantini, Fabrizio Marozzo, Alessio Orsino, Domenico Talia and Paolo Trunfio
Future Internet 2021, 13(5), 121; https://doi.org/10.3390/fi13050121 - 5 May 2021
Cited by 4 | Viewed by 2728
Abstract
Workflows are largely used to orchestrate complex sets of operations required to handle and process huge amounts of data. Parallel processing is often vital to reduce execution time when complex data-intensive workflows must be run efficiently, and at the same time, in-memory processing [...] Read more.
Workflows are largely used to orchestrate complex sets of operations required to handle and process huge amounts of data. Parallel processing is often vital to reduce execution time when complex data-intensive workflows must be run efficiently, and at the same time, in-memory processing can bring important benefits to accelerate execution. However, optimization techniques are necessary to fully exploit in-memory processing, avoiding performance drops due to memory saturation events. This paper proposed a novel solution, called the Intelligent In-memory Workflow Manager (IIWM), for optimizing the in-memory execution of data-intensive workflows on parallel machines. IIWM is based on two complementary strategies: (1) a machine learning strategy for predicting the memory occupancy and execution time of workflow tasks; (2) a scheduling strategy that allocates tasks to a computing node, taking into account the (predicted) memory occupancy and execution time of each task and the memory available on that node. The effectiveness of the machine learning-based predictor and the scheduling strategy were demonstrated experimentally using as a testbed, Spark, a high-performance Big Data processing framework that exploits in-memory computing to speed up the execution of large-scale applications. In particular, two synthetic workflows were prepared for testing the robustness of the IIWM in scenarios characterized by a high level of parallelism and a limited amount of memory reserved for execution. Furthermore, a real data analysis workflow was used as a case study, for better assessing the benefits of the proposed approach. Thanks to high accuracy in predicting resources used at runtime, the IIWM was able to avoid disk writes caused by memory saturation, outperforming a traditional strategy in which only dependencies among tasks are taken into account. Specifically, the IIWM achieved up to a 31% and a 40% reduction of makespan and a performance improvement up to 1.45× and 1.66× on the synthetic workflows and the real case study, respectively. Full article
(This article belongs to the Special Issue The Future of Supercomputing)
►▼ Show Figures

Figure 1

Figure 1
<p>Correlation of target variables with the other features.</p> Full article ">Figure 2
<p>Meta learner regression estimates for the different target variables.</p> Full article ">Figure 3
<p>Execution flow of the IIWM scheduler. Given a workflow and a prediction model as the input, a scheduling plan is generated in two steps: (i) building of the stages and task assignment; (ii) stage consolidation.</p> Full article ">Figure 4
<p>Task dependencies (Workflow 1).</p> Full article ">Figure 5
<p>Disk usage over time for full-parallel and the IIWM.</p> Full article ">Figure 6
<p>Average peak disk usage and execution time, varying the size of available memory.</p> Full article ">Figure 7
<p>Task dependencies (Workflow 2).</p> Full article ">Figure 8
<p>Disk usage over time for full-parallel and the IIWM.</p> Full article ">Figure 9
<p>Ensemble learning workflow.</p> Full article ">Figure 10
<p>Disk usage over time for full-parallel and the IIWM.</p> Full article ">Figure 11
<p>Average peak disk usage and execution time, varying the size of available memory.</p> Full article ">
16 pages, 2846 KiB  
Article
Analysis and Prediction of “AI + Education” Attention Based on Baidu Index—Taking Guizhou Province as an Example
by Yulin Zhao, Junke Li and Jiang-E Wang
Future Internet 2021, 13(5), 120; https://doi.org/10.3390/fi13050120 - 30 Apr 2021
Cited by 8 | Viewed by 3505
Abstract
Studying the attention of “artificial intelligence + education” in ethnic areas is of great significance for China for promoting the integrated development of new educational modes and modern technology in the western region. Guizhou province is an area inhabited by ethnic minorities, located [...] Read more.
Studying the attention of “artificial intelligence + education” in ethnic areas is of great significance for China for promoting the integrated development of new educational modes and modern technology in the western region. Guizhou province is an area inhabited by ethnic minorities, located in the heart of Southwest China. The development of its intelligent education has strong enlightenment for the whole country and the region. Therefore, this paper selects the Baidu Index of “artificial intelligence (AI) + education” in Guizhou province from 2013 to 2020, analyzes the spatial–temporal characteristics of its network attention by using the elastic coefficient method, and builds the ARIMA model on this basis to predict future development. The results show that the public’s attention to “AI + education” differs significantly in time and space. Then, according to the prediction results, this paper puts forward relevant suggestions for the country to promote the sustainable development of education in western ethnic areas. Full article
(This article belongs to the Special Issue Education 4.0 in the Transformation of Universities: Educational and Research Applications)
►▼ Show Figures

Figure 1

Figure 1
<p>Evolution trend of network attention of “AI + education” from 2013 to 2020.</p> Full article ">Figure 2
<p>Change trend of the elasticity coefficient of attention of “AI + education” in Guizhou province from 2014 to 2019.</p> Full article ">Figure 3
<p>Geographical location distribution map of various cities (prefectures) in Guizhou province.</p> Full article ">Figure 4
<p>The mean change trend of “AI + education” network attention in various cities (prefectures) of Guizhou province from 2013 to 2020.</p> Full article ">Figure 5
<p>Total Baidu search index value of “AI + education” in Guizhou province.</p> Full article ">Figure 6
<p>Sequence after differential treatment.</p> Full article ">Figure 7
<p>Mean and variance of the sequence of second-order difference.</p> Full article ">Figure 8
<p>Fit diagram of the original value and predicted value of attention to “AI + education” in Guizhou province in a period.</p> Full article ">Figure 9
<p>Forecast trend of “AI + education” attention in the cities (prefectures) of Guizhou province.</p> Full article ">Figure 10
<p>Forecast trend of “AI + education” attention in other cities.</p> Full article ">
17 pages, 1410 KiB  
Article
The Social and Transfer Massive Open Online Course: Post-Digital Learning
by José Javier Hueso-Romero, Javier Gil-Quintana, Helen Hasbun and Sara Osuna-Acedo
Future Internet 2021, 13(5), 119; https://doi.org/10.3390/fi13050119 - 30 Apr 2021
Cited by 10 | Viewed by 3694
Abstract
This research provides a current view on post-digital learning experiences with a massive open online course (MOOC), in relation to user profiles, universal instructional design, digital resources, inclusive activities and collaborative assessment. The study is based on a mixed research methodology, creating a [...] Read more.
This research provides a current view on post-digital learning experiences with a massive open online course (MOOC), in relation to user profiles, universal instructional design, digital resources, inclusive activities and collaborative assessment. The study is based on a mixed research methodology, creating a questionnaire aimed at people with experience in any MOOC typology, in which the learning methodology, the instructional didactic design of the MOOCs, the resources, proposed activities, and accessibility are analyzed. Additionally, interviews and focus groups were carried out with the creators of massive open online social courses alongside the students of the official Master of Communication and Education on the Internet, offered by the UNED (Universidad Nacional de Educación a Distancia—Spain), with the subject of virtual participation scenarios. The data obtained are subjected to statistical tests to determine the scientific rigor, such as Cronbach’s alpha, the Kolmogorov-Smirnov normality test, and the non-parametric tests of Spearman’s correlation coefficient and Kendall’s Tau b. In conclusion, the social massive open online course/transfer massive open online course model is evidenced as a projected approach in social networks. The sMOOC and tMOOC are online training models, which are in constant development and evolution, as a social, creative, collaborative, interactive, and inclusive learning methodology, offering new challenges for the digital distance education of the future. The research carried out is only related and linked to the experiences of different people with the sMOOC and tMOOC. Full article
(This article belongs to the Special Issue Social Network and Sustainable Distance Education)
►▼ Show Figures

Figure 1

Figure 1
<p>Respondent profile. Identification questions.</p> Full article ">Figure 2
<p>Frequency distribution of the variable total_items.</p> Full article ">Figure 3
<p>Most used platforms in the experiences with MOOC typologies.</p> Full article ">Figure 4
<p>Mode of evaluation in MOOC experiences.</p> Full article ">
23 pages, 700 KiB  
Review
Survey on Intelligence Edge Computing in 6G: Characteristics, Challenges, Potential Use Cases, and Market Drivers
by Ahmed Al-Ansi, Abdullah M. Al-Ansi, Ammar Muthanna, Ibrahim A. Elgendy and Andrey Koucheryavy
Future Internet 2021, 13(5), 118; https://doi.org/10.3390/fi13050118 - 30 Apr 2021
Cited by 87 | Viewed by 9670
Abstract
Intelligence Edge Computing (IEC) is the key enabler of emerging 5G technologies networks and beyond. IEC is considered to be a promising backbone of future services and wireless communication systems in 5G integration. In addition, IEC enables various use cases and applications, including [...] Read more.
Intelligence Edge Computing (IEC) is the key enabler of emerging 5G technologies networks and beyond. IEC is considered to be a promising backbone of future services and wireless communication systems in 5G integration. In addition, IEC enables various use cases and applications, including autonomous vehicles, augmented and virtual reality, big data analytic, and other customer-oriented services. Moreover, it is one of the 5G technologies that most enhanced market drivers in different fields such as customer service, healthcare, education methods, IoT in agriculture and energy sustainability. However, 5G technological improvements face many challenges such as traffic volume, privacy, security, digitization capabilities, and required latency. Therefore, 6G is considered to be promising technology for the future. To this end, compared to other surveys, this paper provides a comprehensive survey and an inclusive overview of Intelligence Edge Computing (IEC) technologies in 6G focusing on main up-to-date characteristics, challenges, potential use cases and market drivers. Furthermore, we summarize research efforts on IEC in 5G from 2014 to 2021, in which the integration of IEC and 5G technologies are highlighted. Finally, open research challenges and new future directions in IEC with 6G networks will be discussed. Full article
(This article belongs to the Special Issue 6G Wireless Communication Systems: Applications, Opportunities and Challenges)
►▼ Show Figures

Figure 1

Figure 1
<p>Vision of 6G.</p> Full article ">Figure 2
<p>IEC Architecture [<a href="#B9-futureinternet-13-00118" class="html-bibr">9</a>].</p> Full article ">Figure 3
<p>IEC and 6G Integration.</p> Full article ">
29 pages, 1308 KiB  
Article
H2O: Secure Interactions in IoT via Behavioral Fingerprinting
by Marco Ferretti, Serena Nicolazzo and Antonino Nocera
Future Internet 2021, 13(5), 117; https://doi.org/10.3390/fi13050117 - 30 Apr 2021
Cited by 10 | Viewed by 2805
Abstract
Sharing data and services in the Internet of Things (IoT) can give rise to significant security concerns with information being sensitive and vulnerable to attacks. In such an environment, objects can be either public resources or owned by humans. For this reason, the [...] Read more.
Sharing data and services in the Internet of Things (IoT) can give rise to significant security concerns with information being sensitive and vulnerable to attacks. In such an environment, objects can be either public resources or owned by humans. For this reason, the need of monitoring the reliability of all involved actors, both persons and smart objects, assuring that they really are who they claim to be, is becoming an essential property of the IoT, with the increase in the pervasive adoption of such a paradigm. In this paper, we tackle this problem by proposing a new framework, called H2O (Human to Object). Our solution is able to continuously authenticate an entity in the network, providing a reliability assessment mechanism based on behavioral fingerprinting. A detailed security analysis evaluates the robustness of the proposed protocol; furthermore, a performance analysis shows the feasibility of our approach. Full article
(This article belongs to the Special Issue Human–Computer Interaction Models and Experiences for Internet of Things Systems and Edge Computing)
►▼ Show Figures

Figure 1

Figure 1
<p>A representation different nodes’ groups in H2O Network.</p> Full article ">Figure 2
<p>A representation of H2O Network with entity operations.</p> Full article ">Figure 3
<p>The steps of our approach for a human-to-human interaction.</p> Full article ">Figure 4
<p>A network portion of H2O.</p> Full article ">Figure 5
<p>The sequence diagram for a human-to-human communication.</p> Full article ">Figure 6
<p>The maps of smart things of Santander used in our experiments.</p> Full article ">Figure 7
<p>Different percentage of transaction between our solution and a naive approach.</p> Full article ">Figure 8
<p>Nodes degree distribution in H2O network.</p> Full article ">Figure 9
<p>The percentage of assessable nodes for different values of <span class="html-italic">r</span>.</p> Full article ">Figure 10
<p>Number of prevented attacks against the number of hacked/corrupted nodes with different values of <span class="html-italic">r</span>.</p> Full article ">
15 pages, 1896 KiB  
Article
A Framework for Mobile-Assisted Formative Assessment to Promote Students’ Self-Determination
by Stavros A. Nikou and Anastasios A. Economides
Future Internet 2021, 13(5), 116; https://doi.org/10.3390/fi13050116 - 30 Apr 2021
Cited by 15 | Viewed by 4271
Abstract
Motivation is an important issue to consider when designing learning activities, including mobile learning and assessment. While previous research provides evidence for the motivational impact of mobile learning, not many pedagogical frameworks exist for the design of mobile-assisted learning and assessment. The current [...] Read more.
Motivation is an important issue to consider when designing learning activities, including mobile learning and assessment. While previous research provides evidence for the motivational impact of mobile learning, not many pedagogical frameworks exist for the design of mobile-assisted learning and assessment. The current study is grounded in the Self-Determination Theory of motivation and proposes a pedagogical framework for mobile-assisted formative assessment, aiming at enhancing student motivation. For a preliminary evaluation of the framework, fifty-one students from a public European high school participated in a series of formative assessment activities. The tasks that were implemented according to the proposed mobile-based formative assessment framework had a significant positive impact on student perceived levels of autonomy, competence, and relatedness, enhancing students’ intrinsic motivation levels. Study findings highlighted the capacity of the proposed framework to guide the design of mobile-based formative assessment activities that enhance and promote student motivation. The study makes a theoretical contribution by proposing a framework that aligns mobile learning and assessment with elements of the Self-Determination Theory of motivation and also has a practical contribution by implementing mobile learning and assessment practices that have the potential to promote student motivation. Full article
(This article belongs to the Special Issue Technology Enhanced Learning and Mobile Learning)
►▼ Show Figures

Figure 1

Figure 1
<p>Formative assessment framework to enhance student motivation using mobile technologies (notation explained: (a) stands for autonomy, (c) for competence, and (r) for relatedness support, e.g., (a1.1) supports autonomy, (cr2.3) supports competence and relatedness).</p> Full article ">Figure 2
<p>Motivation supportive pedagogies.</p> Full article ">Figure 3
<p>Motivation supportive mobile technologies.</p> Full article ">Figure 4
<p>The experimental procedure.</p> Full article ">Figure 5
<p>Implemented components of the framework.</p> Full article ">Figure 6
<p>Screenshots of the mobile-based formative assessment.</p> Full article ">Figure 7
<p>Students participating in the formative assessment activities.</p> Full article ">Figure 8
<p>Perceived pre- and post-levels of autonomy, competence, and relatedness.</p> Full article ">
22 pages, 11535 KiB  
Article
Experimental Evaluation of a LoRa Wildlife Monitoring Network in a Forest Vegetation Area
by Mike Oluwatayo Ojo, Davide Adami and Stefano Giordano
Future Internet 2021, 13(5), 115; https://doi.org/10.3390/fi13050115 - 29 Apr 2021
Cited by 48 | Viewed by 6309
Abstract
Smart agriculture and wildlife monitoring are one of the recent trends of Internet of Things (IoT) applications, which are evolving in providing sustainable solutions from producers. This article details the design, development and assessment of a wildlife monitoring application for IoT animal repelling [...] Read more.
Smart agriculture and wildlife monitoring are one of the recent trends of Internet of Things (IoT) applications, which are evolving in providing sustainable solutions from producers. This article details the design, development and assessment of a wildlife monitoring application for IoT animal repelling devices that is able to cover large areas, thanks to the low power wide area networks (LPWAN), which bridge the gap between cellular technologies and short range wireless technologies. LoRa, the global de-facto LPWAN, continues to attract attention given its open specification and ready availability of off-the-shelf hardware, with claims of several kilometers of range in harsh challenging environments. At first, this article presents a survey of the LPWAN for smart agriculture applications. We proceed to evaluate the performance of LoRa transmission technology operating in the 433 MHz and 868 MHz bands, aimed at wildlife monitoring in a forest vegetation area. To characterize the communication link, we mainly use the signal-to-noise ratio (SNR), received signal strength indicator (RSSI) and packet delivery ratio (PDR). Findings from this study show that achievable performance can greatly vary between the 433 MHz and 868 MHz bands, and prompt caution is required when taking numbers at face value, as this can have implications for IoT applications. In addition, our results show that the link reaches up to 860 m in the highly dense forest vegetation environment, while in the not so dense forest vegetation environment, it reaches up to 2050 m. Full article
►▼ Show Figures

Figure 1

Figure 1
<p>Low Power Wide Area Network Classification.</p> Full article ">Figure 2
<p>LoRa Frame Structure.</p> Full article ">Figure 3
<p>System Architecture.</p> Full article ">Figure 4
<p>Repeller board sketch (<b>a</b>) wide open (<b>b</b>) in an enclosure.</p> Full article ">Figure 5
<p>LoRa Field Packet Generator.</p> Full article ">Figure 6
<p>RN2483 Block Diagram.</p> Full article ">Figure 7
<p>The Prototype devices (LoRa end-node) placed at different positions (POS A–POS F, POS 1–POS 4), and the Gateway indicated as POSRX.</p> Full article ">Figure 8
<p>Image showing the environment in the green and red subsets.</p> Full article ">Figure 9
<p>(<b>a</b>) PDR vs. payload length for SF12, CR = 4/5, BW = 125 KHz @ 868 MHz (<b>b</b>) log(PDR) vs. payload length for SF12.</p> Full article ">Figure 10
<p>PDR vs. SF in POS 2 with CR = 4/5, BW = 125 KHz for (<b>a</b>) 433 MHz band (<b>b</b>) 868 MHz band.</p> Full article ">Figure 11
<p>PDR vs. SF varying the CR in POS C (<b>a</b>) 433 MHz (<b>b</b>) 868 MHz.</p> Full article ">Figure 12
<p>PDR vs. SF varying the CR in POS F (<b>a</b>) 433 MHz (<b>b</b>) 868 MHz.</p> Full article ">Figure 13
<p>PDR vs. SF varying the CR in POS 3 (<b>a</b>) 433 MHz (<b>b</b>) 868 MHz.</p> Full article ">Figure 14
<p>RSSI registered between the GW and POS 2 for different SF and CR values.</p> Full article ">Figure 15
<p>PDR achieved for transmissions from all positions for different SF values, with CR = 4/5, BW = 125 KHz in the 433 MHz and 868 MHz frequency bands.</p> Full article ">Figure 16
<p>SNR vs. SF with CR = 4/5, BW = 125 KHz for POS 1–POS 3.</p> Full article ">Figure 17
<p>Theoretical coverage heat map for the forest environment area (<b>a</b>) 433 MHz band (<b>b</b>) 868 MHz band.</p> Full article ">
25 pages, 3390 KiB  
Article
Collecting a Large Scale Dataset for Classifying Fake News Tweets Using Weak Supervision
by Stefan Helmstetter and Heiko Paulheim
Future Internet 2021, 13(5), 114; https://doi.org/10.3390/fi13050114 - 29 Apr 2021
Cited by 25 | Viewed by 4725
Abstract
The problem of automatic detection of fake news in social media, e.g., on Twitter, has recently drawn some attention. Although, from a technical perspective, it can be regarded as a straight-forward, binary classification problem, the major challenge is the collection of large enough [...] Read more.
The problem of automatic detection of fake news in social media, e.g., on Twitter, has recently drawn some attention. Although, from a technical perspective, it can be regarded as a straight-forward, binary classification problem, the major challenge is the collection of large enough training corpora, since manual annotation of tweets as fake or non-fake news is an expensive and tedious endeavor, and recent approaches utilizing distributional semantics require large training corpora. In this paper, we introduce an alternative approach for creating a large-scale dataset for tweet classification with minimal user intervention. The approach relies on weak supervision and automatically collects a large-scale, but very noisy, training dataset comprising hundreds of thousands of tweets. As a weak supervision signal, we label tweets by their source, i.e., trustworthy or untrustworthy source, and train a classifier on this dataset. We then use that classifier for a different classification target, i.e., the classification of fake and non-fake tweets. Although the labels are not accurate according to the new classification target (not all tweets by an untrustworthy source need to be fake news, and vice versa), we show that despite this unclean, inaccurate dataset, the results are comparable to those achieved using a manually labeled set of tweets. Moreover, we show that the combination of the large-scale noisy dataset with a human labeled one yields more advantageous results than either of the two alone. Full article
(This article belongs to the Special Issue Digital and Social Media in the Disinformation Age)
►▼ Show Figures

Figure 1

Figure 1
<p>Distribution of tweets labeled as real and fake news in the training dataset.</p> Full article ">Figure 2
<p>Topic distribution in the training dataset.</p> Full article ">Figure 3
<p>Statistics on sentiment and subjectivity score, as well as on retweet and favorite counts.</p> Full article ">Figure 4
<p>Schematic depiction of the overall approach.</p> Full article ">Figure 5
<p>F1 scores using different dimensions for Doc2vec, using DBOW (in blue) and DM (in red).</p> Full article ">Figure 6
<p>F1 scores using different numbers of topics for LDA with different learning algorithms (in blue), and the results for HDP (in red).</p> Full article ">Figure 7
<p>Representation of polarity and subjectivity scores by [<a href="#B43-futureinternet-13-00114" class="html-bibr">43</a>].</p> Full article ">Figure 8
<p>Feature importance for Scenario 1 (tweet level features only), as computed by XGBoost.</p> Full article ">Figure 9
<p>Feature importance for Scenario 2 (tweet and user level features), as computed by XGBoost.</p> Full article ">
15 pages, 1138 KiB  
Article
Dynamic Control Architecture Based on Software Defined Networking for the Internet of Things
by Michele Bonanni, Francesco Chiti, Romano Fantacci and Laura Pierucci
Future Internet 2021, 13(5), 113; https://doi.org/10.3390/fi13050113 - 28 Apr 2021
Cited by 14 | Viewed by 3174
Abstract
Software Defined Networking (SDN) provides a new perspective for the Internet of Things (IoT), since, with the separation of the control from the data planes, it is viable to optimise the traditional networks operation management. In particular, the SDN Controller has a global [...] Read more.
Software Defined Networking (SDN) provides a new perspective for the Internet of Things (IoT), since, with the separation of the control from the data planes, it is viable to optimise the traditional networks operation management. In particular, the SDN Controller has a global vision of the network of sensors/actuators domain, allowing real-time network nodes and data flows reconfiguration. As a consequence, devices, usually facing limited communications and computing resources, are relieved of the route selection task in a distributed and, thus, suboptimal way. This paper proposes a SDN-IoT architecture, specifically focusing on the Controller design, which dynamically optimises in real time the end-to-end flows delivery. In particular, the dynamic routing policy adaptation is based on the real-time estimation of the network status and it allows jointly minimising the end-to-end latency and energy consumption and, consequently, to improve the network life time. The performance of the proposed approach is analysed in terms of the average latency, energy consumption and overhead, pointing out a better behaviour in comparison with the existing distributed approaches. Full article
(This article belongs to the Special Issue Service-Oriented Systems and Applications)
►▼ Show Figures

Figure 1

Figure 1
<p>Reference SDN-IoT architecture, which shows the presence of Application, Control and Data Planes. AP is comprised of both local and remote applications, while CP is composed of a Controller with endowed the routing module (DRM). Finally, generic sensor nodes are within DP, where only Root is physically connected to the Controller.</p> Full article ">Figure 2
<p>Bottleneck topology with 50 randomly deployed nodes.</p> Full article ">Figure 3
<p>Average Residual Energy achieved by the proposed strategies MaxBattery, DRM and MinHop.</p> Full article ">Figure 4
<p>Jain’s Fairness Index for the proposed strategies: MaxBattery, DRM and MinHop.</p> Full article ">Figure 5
<p>Comparison of proposed strategies MaxBattery, DRM and MinHop in terms of Average Latency.</p> Full article ">Figure 6
<p>Overhead comparison for the proposed strategies: MaxBattery, DRM and MinHop.</p> Full article ">Figure 7
<p>Average latency comparison for different network size.</p> Full article ">Figure 8
<p>Average energy consumption comparison.</p> Full article ">Figure 9
<p>Objective function analysis.</p> Full article ">
24 pages, 24633 KiB  
Article
Exploring the Roles of Local Mobility Patterns, Socioeconomic Conditions, and Lockdown Policies in Shaping the Patterns of COVID-19 Spread
by Mauricio Herrera and Alex Godoy-Faúndez
Future Internet 2021, 13(5), 112; https://doi.org/10.3390/fi13050112 - 28 Apr 2021
Cited by 7 | Viewed by 3546
Abstract
The COVID-19 crisis has shown that we can only prevent the risk of mass contagion through timely, large-scale, coordinated, and decisive actions. This pandemic has also highlighted the critical importance of generating rigorous evidence for decision-making, and actionable insights from data, considering further [...] Read more.
The COVID-19 crisis has shown that we can only prevent the risk of mass contagion through timely, large-scale, coordinated, and decisive actions. This pandemic has also highlighted the critical importance of generating rigorous evidence for decision-making, and actionable insights from data, considering further the intricate web of causes and drivers behind observed patterns of contagion diffusion. Using mobility, socioeconomic, and epidemiological data recorded throughout the pandemic development in the Santiago Metropolitan Region, we seek to understand the observed patterns of contagion. We characterize human mobility patterns during the pandemic through different mobility indices and correlate such patterns with the observed contagion diffusion, providing data-driven models for insights, analysis, and inferences. Through these models, we examine some effects of the late application of mobility restrictions in high-income urban regions that were affected by high contagion rates at the beginning of the pandemic. Using augmented synthesis control methods, we study the consequences of the early lifting of mobility restrictions in low-income sectors connected by public transport to high-risk and high-income communes. The Santiago Metropolitan Region is one of the largest Latin American metropolises with features that are common to large cities. Therefore, it can be used as a relevant case study to unravel complex patterns of the spread of COVID-19. Full article
(This article belongs to the Section Techno-Social Smart Systems)
►▼ Show Figures

Figure 1

Figure 1
<p>Capturing mobility patterns of MR through three sensing methodologies. Panel (<b>A</b>) shows around 12,000 bus stations in the MR public transport network. The O-D matrices are estimated by the flow of passengers between these bus stations. Panel (<b>B</b>) indicates the network of cell phone antennas that provide coverage in the MR. The mobility index is determined by the flow of mobile phone users and the records they leave on the antennas covering the service while they are moving. Panel (<b>C</b>) shows level-6 hexagons covering the MR on a given day. The highest scores/percentile are in blue, while the lowest is yellow, passing through a range of greens.</p> Full article ">Figure 2
<p>Time series for the average mobility scores of a sample of 16 Metropolitan Region communes.</p> Full article ">Figure 3
<p>Hierarchical clustering using DTW for mobility time series. The solution with five clusters is highlighted.</p> Full article ">Figure 4
<p>The time series of the commune mobility scores that belong to the five cluster solution. (<b>1</b>) <span class="html-italic">Cluster</span><math display="inline"><semantics> <msub> <mrow/> <mn>1</mn> </msub> </semantics></math>; (<b>2</b>) <span class="html-italic">Cluster</span><math display="inline"><semantics> <msub> <mrow/> <mn>2</mn> </msub> </semantics></math>; (<b>3</b>) <span class="html-italic">Cluster</span><math display="inline"><semantics> <msub> <mrow/> <mn>3</mn> </msub> </semantics></math>; (<b>4</b>) <span class="html-italic">Cluster</span><math display="inline"><semantics> <msub> <mrow/> <mn>4</mn> </msub> </semantics></math>; (<b>5</b>) <span class="html-italic">Cluster</span><math display="inline"><semantics> <msub> <mrow/> <mn>5</mn> </msub> </semantics></math>.</p> Full article ">Figure 5
<p>A sample of 10 communes grouped in clusters of similar behavior according to the values of the scores. The central communes of the MR (Ñuñoa, Santiago, Estación Central) have the highest scores, while the peripheral rural communes (Paine, Pirque, Tiltil) have the lowest scores. It is also observed that communes with very different incomes may have similar mobility scores (low-income commune: Puente Alto, and high-income communes: VItacura, Las Condes).</p> Full article ">Figure 6
<p>Time series of the mobility index for a sample of 16 MR communes in the period between 27 February 2020 and 5 November 2020. The dates corresponding to the most important confinement measures are highlighted.</p> Full article ">Figure 7
<p>Representation of the correlations between the daily values of the mobility indices for the communes of the Metropolitan Region. A high positive correlation is observed in the mobility of all the communes.</p> Full article ">Figure 8
<p>Origin–destination between communes of Santiago City (<b>A</b>) at 6:00 a.m. and (<b>B</b>) at 6:00 p.m. on a regular pre-pandemic weekday. The sizes of the arrows are related to the number of trips, and the sizes of the nodes with its out-degrees.</p> Full article ">Figure 9
<p>Cumulative cases of COVID-19 for the communes of the Metropolitan Region.</p> Full article ">Figure 10
<p>Temporal evolution of the correlation between the mobility index and the cumulative number of COVID-19 cases per 100.000 inhabitants in the period from 26 March to 30 October 2020.</p> Full article ">Figure 11
<p>Mobility index vs. cumulative number of cases per 100.000 inhabitants of the commune at the beginning of the total quarantine for the communes of the Metropolitan Region. The radii of the bubbles are proportional to the commune population.</p> Full article ">Figure 12
<p>The mobility index vs. the cumulative number of cases per 100.000 inhabitants after relaxing the confinement measures.</p> Full article ">Figure 13
<p>Estimation of the effect of lifting the quarantine on COVID-19 new cases for low-income communes in the Metropolitan Region.</p> Full article ">Figure 14
<p>Estimation of the effect of lifting the quarantine on mobility in low-income communes of the Metropolitan Region.</p> Full article ">
18 pages, 516 KiB  
Article
Designing a Network Intrusion Detection System Based on Machine Learning for Software Defined Networks
by Abdulsalam O. Alzahrani and Mohammed J. F. Alenazi
Future Internet 2021, 13(5), 111; https://doi.org/10.3390/fi13050111 - 28 Apr 2021
Cited by 165 | Viewed by 12249
Abstract
Software-defined Networking (SDN) has recently developed and been put forward as a promising and encouraging solution for future internet architecture. Managed, the centralized and controlled network has become more flexible and visible using SDN. On the other hand, these advantages bring us a [...] Read more.
Software-defined Networking (SDN) has recently developed and been put forward as a promising and encouraging solution for future internet architecture. Managed, the centralized and controlled network has become more flexible and visible using SDN. On the other hand, these advantages bring us a more vulnerable environment and dangerous threats, causing network breakdowns, systems paralysis, online banking frauds and robberies. These issues have a significantly destructive impact on organizations, companies or even economies. Accuracy, high performance and real-time systems are essential to achieve this goal successfully. Extending intelligent machine learning algorithms in a network intrusion detection system (NIDS) through a software-defined network (SDN) has attracted considerable attention in the last decade. Big data availability, the diversity of data analysis techniques, and the massive improvement in the machine learning algorithms enable the building of an effective, reliable and dependable system for detecting different types of attacks that frequently target networks. This study demonstrates the use of machine learning algorithms for traffic monitoring to detect malicious behavior in the network as part of NIDS in the SDN controller. Different classical and advanced tree-based machine learning techniques, Decision Tree, Random Forest and XGBoost are chosen to demonstrate attack detection. The NSL-KDD dataset is used for training and testing the proposed methods; it is considered a benchmarking dataset for several state-of-the-art approaches in NIDS. Several advanced preprocessing techniques are performed on the dataset in order to extract the best form of the data, which produces outstanding results compared to other systems. Using just five out of 41 features of NSL-KDD, a multi-class classification task is conducted by detecting whether there is an attack and classifying the type of attack (DDoS, PROBE, R2L, and U2R), accomplishing an accuracy of 95.95%. Full article
(This article belongs to the Special Issue Mobile and Wireless Network Security and Privacy)
►▼ Show Figures

Figure 1

Figure 1
<p>Proposed System Architecture consists of three main layers: the application layer, which contains the Rule management system and NIDS; the control layer contains the SDN controller, while the infrastructure layer includes the switches, servers and the attacker that may exist in one of those switches.</p> Full article ">Figure 2
<p>Proposed System Architecture. This flow chart represents the overall process. The process starts with the data source, the NSL-KDD dataset, followed by data analysis and some pre-processing techniques for data cleaning. Feature normalization and feature selection are performed in order to implement the later selected tree-based machine learning algorithm to classify whether there is an attack or not and the types of attack.</p> Full article ">Figure 3
<p>Categorized training set statistics show the number of samples in the four types of attacks and the normal samples in the training set.</p> Full article ">Figure 4
<p>Categorized test set statistics show the number of samples in the four types of attacks and the normal samples in the test set. These samples are an unseen set, for model evaluation.</p> Full article ">Figure 5
<p>The proposed model XGBoost is compared with other tree-based machine learning algorithms, using F1-score, precision and recall as evaluation metrics.</p> Full article ">Figure 6
<p>Accuracy score is measured in order to evaluate the performance of the proposed XGBoost model against the Deep Neural Networks model.</p> Full article ">Figure 7
<p>Performance evaluation of the proposed model in detecting attacks for each category, using F-score, precision and recall.</p> Full article ">Figure 8
<p>Computing the ROC curve by measuring the true-positive rate against the false-positive rate of the proposed model on each attack category (DOS, Probe, R2L, and U2R).</p> Full article ">Figure 9
<p>Attack detection rate comparison between the proposed XGboost method and [<a href="#B27-futureinternet-13-00111" class="html-bibr">27</a>].</p> Full article ">Figure 10
<p>Accuracy performance comparison for proposed XGboost method against classical machine learning algorithms, including the DNN model in [<a href="#B27-futureinternet-13-00111" class="html-bibr">27</a>].</p> Full article ">
12 pages, 558 KiB  
Article
Reviewing Stranger on the Internet: The Role of Identifiability through “Reputation” in Online Decision Making
by Mirko Duradoni, Stefania Collodi, Serena Coppolino Perfumi and Andrea Guazzini
Future Internet 2021, 13(5), 110; https://doi.org/10.3390/fi13050110 - 27 Apr 2021
Cited by 12 | Viewed by 3809
Abstract
The stranger on the Internet effect has been studied in relation to self-disclosure. Nonetheless, quantitative evidence about how people mentally represent and perceive strangers online is still missing. Given the dynamic development of web technologies, quantifying how much strangers can be considered suitable [...] Read more.
The stranger on the Internet effect has been studied in relation to self-disclosure. Nonetheless, quantitative evidence about how people mentally represent and perceive strangers online is still missing. Given the dynamic development of web technologies, quantifying how much strangers can be considered suitable for pro-social acts such as self-disclosure appears fundamental for a whole series of phenomena ranging from privacy protection to fake news spreading. Using a modified and online version of the Ultimatum Game (UG), we quantified the mental representation of the stranger on the Internet effect and tested if people modify their behaviors according to the interactors’ identifiability (i.e., reputation). A total of 444 adolescents took part in a 2 × 2 design experiment where reputation was set active or not for the two traditional UG tasks. We discovered that, when matched with strangers, people donate the same amount of money as if the other has a good reputation. Moreover, reputation significantly affected the donation size, the acceptance rate and the feedback decision making as well. Full article
(This article belongs to the Special Issue Selected Papers from the INSCI2019: Internet Science 2019)
►▼ Show Figures

Figure 1

Figure 1
<p>Acceptance dynamics with respect to positive and negative reputation of the interactor. In the insert, the acceptance rate trend related to the receiving conditions is represented.</p> Full article ">Figure 2
<p>Feedback coherence trend in relation to Donors’ reputation levels.</p> Full article ">Figure 3
<p>Comparison between the average donation trend with respect to Receivers’ reputation scores (dots) and the amount offered without the reputation system (dashed line).</p> Full article ">
12 pages, 3467 KiB  
Article
How Schools Affected the COVID-19 Pandemic in Italy: Data Analysis for Lombardy Region, Campania Region, and Emilia Region
by Davide Tosi and Alessandro Siro Campi
Future Internet 2021, 13(5), 109; https://doi.org/10.3390/fi13050109 - 27 Apr 2021
Cited by 11 | Viewed by 6637
Abstract
Background: Coronavirus Disease 2019 (COVID-19) is the main discussed topic worldwide in 2020 and at the beginning of the Italian epidemic, scientists tried to understand the virus diffusion and the epidemic curve of positive cases with controversial findings and numbers. Objectives: In this [...] Read more.
Background: Coronavirus Disease 2019 (COVID-19) is the main discussed topic worldwide in 2020 and at the beginning of the Italian epidemic, scientists tried to understand the virus diffusion and the epidemic curve of positive cases with controversial findings and numbers. Objectives: In this paper, a data analytics study on the diffusion of COVID-19 in Lombardy Region and Campania Region is developed in order to identify the driver that sparked the second wave in Italy. Methods: Starting from all the available official data collected about the diffusion of COVID-19, we analyzed Google mobility data, school data and infection data for two big regions in Italy: Lombardy Region and Campania Region, which adopted two different approaches in opening and closing schools. To reinforce our findings, we also extended the analysis to the Emilia Romagna Region. Results: The paper shows how different policies adopted in school opening/closing may have had an impact on the COVID-19 spread, while other factors related to citizen mobility did not affect the second Italian wave. Conclusions: The paper shows that a clear correlation exists between the school contagion and the subsequent temporal overall contagion in a geographical area. Moreover, it is clear that highly populated provinces have the greatest spread of the virus. Full article
(This article belongs to the Special Issue Software Engineering and Data Science)
►▼ Show Figures

Figure 1

Figure 1
<p>Observed real curves for new daily cases and deaths in Italy.</p> Full article ">Figure 2
<p>Lombardy Region data set for school contagion index and overall contagion. Correlation study.</p> Full article ">Figure 3
<p>Lombardy Region scatter plot for school contagion index vs. overall contagion index.</p> Full article ">Figure 4
<p>Campania Region data set for school contagion index and overall contagion. Correlation study.</p> Full article ">Figure 4 Cont.
<p>Campania Region data set for school contagion index and overall contagion. Correlation study.</p> Full article ">Figure 5
<p>Emilia Romagna Region data set for school contagion index and overall contagion. Correlation study.</p> Full article ">Figure 6
<p>Emilia Romagna scatter plot for school contagion index vs. overall contagion index.</p> Full article ">Figure 7
<p>Daily positive new cases and 7-days avg. trend for the Italian regions: Campania, Emilia and Lombardy in the time frame 1 September 2020 to 15 November 2020.</p> Full article ">Figure 8
<p>Rt trend for the three Italian regions: Campania, Emilia and Lombardy.</p> Full article ">
45 pages, 10563 KiB  
Article
Inferring Urban Social Networks from Publicly Available Data
by Stefano Guarino, Enrico Mastrostefano, Massimo Bernaschi, Alessandro Celestini, Marco Cianfriglia, Davide Torre and Lena Rebecca Zastrow
Future Internet 2021, 13(5), 108; https://doi.org/10.3390/fi13050108 - 26 Apr 2021
Cited by 12 | Viewed by 5361
Abstract
The definition of suitable generative models for synthetic yet realistic social networks is a widely studied problem in the literature. By not being tied to any real data, random graph models cannot capture all the subtleties of real networks and are inadequate for [...] Read more.
The definition of suitable generative models for synthetic yet realistic social networks is a widely studied problem in the literature. By not being tied to any real data, random graph models cannot capture all the subtleties of real networks and are inadequate for many practical contexts—including areas of research, such as computational epidemiology, which are recently high on the agenda. At the same time, the so-called contact networks describe interactions, rather than relationships, and are strongly dependent on the application and on the size and quality of the sample data used to infer them. To fill the gap between these two approaches, we present a data-driven model for urban social networks, implemented and released as open source software. By using just widely available aggregated demographic and social-mixing data, we are able to create, for a territory of interest, an age-stratified and geo-referenced synthetic population whose individuals are connected by “strong ties” of two types: intra-household (e.g., kinship) or friendship. While household links are entirely data-driven, we propose a parametric probabilistic model for friendship, based on the assumption that distances and age differences play a role, and that not all individuals are equally sociable. The demographic and geographic factors governing the structure of the obtained network, under different configurations, are thoroughly studied through extensive simulations focused on three Italian cities of different size. Full article
(This article belongs to the Section Techno-Social Smart Systems)
►▼ Show Figures

Figure 1

Figure 1
<p>A graphical representation of our territory and population, through a random sample of 1000 people on the territory of Florence and Sabaudia (some points may fall outside of the territory if the center of the tile to which they belong falls inside).</p> Full article ">Figure 2
<p>(<b>Florence</b>) Average degree of the individuals of each age group, in the whole friendship graph <math display="inline"><semantics> <msub> <mi>G</mi> <mi>F</mi> </msub> </semantics></math> and with their peers, under different configurations all with <math display="inline"><semantics> <mrow> <mi>μ</mi> <mo>=</mo> <mn>5</mn> </mrow> </semantics></math>. With respect to the benchmark configuration where all age-groups are equivalent (<b>left panel</b>), the introduction of a data-driven <span class="html-italic">S</span> increases the internal cohesion of younger age groups, but it fails to correctly model inter-group connections (<b>central panel</b>), which are instead well captured if the population is also data-driven (<b>right panel</b>).</p> Full article ">Figure 3
<p>(<b>Florence</b>) Percentage of edges between age groups in a configuration with data-driven population and age-based mixing, for <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math>, <span class="html-italic">μ</span> = 5 and <math display="inline"><semantics> <mrow> <mi>D</mi> <mo stretchy="false">(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo stretchy="false">)</mo> <mo>=</mo> <mi>d</mi> <msup> <mrow> <mo>(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo>)</mo> </mrow> <mrow> <mo>−</mo> <mn>2</mn> </mrow> </msup> </mrow> </semantics></math>. The value in cell <span class="html-italic">i</span>, <span class="html-italic">j</span> is proportional to <span class="html-italic">m<sub>i,j</sub></span>·<span class="html-italic">s<sub>i,j</sub></span>. Adults are still involved, in total, in the majority of friendship edges, even if they have a lower average number of friends than the younger groups.</p> Full article ">Figure 4
<p>(<b>Florence</b>) Degree distributions of friendship graphs with <math display="inline"><semantics> <mrow> <mi>μ</mi> <mo>=</mo> <mn>5</mn> </mrow> </semantics></math> and corresponding Poisson distributions with <math display="inline"><semantics> <mrow> <mi>λ</mi> <mo>=</mo> <mn>5</mn> </mrow> </semantics></math>. The data distribution lines show the mean, whereas the shaded areas around the lines represent the 95% confidence interval. The Kullback–Leibler divergence between the Poisson and the empirical distribution are (left to right, top to bottom): 1.6e-07, 2.87e-07, 1.1e-05, 5.2e-05, 0.0002, 0.004. The degree distribution diverges from a Poisson (which is the expected behavior in ER-like graphs) as we incorporate geographic features in the network, and the effect is especially significant when long edges are strongly penalized.</p> Full article ">Figure 5
<p>(<b>Florence</b>) Degree distributions of friendship graphs with <math display="inline"><semantics> <mrow> <mi>μ</mi> <mo>=</mo> <mn>10</mn> </mrow> </semantics></math> and corresponding Poisson distributions with <math display="inline"><semantics> <mrow> <mi>λ</mi> <mo>=</mo> <mn>10</mn> </mrow> </semantics></math>. The data distribution lines show the mean, whereas the shaded areas around the lines represent the 95% confidence interval. The Kullback–Leibler divergence between the Poisson and the empirical distribution are (left to right, top to bottom): 2.5e-07, 2.01e-07, 3.2e-05, 0.0001, 0.0005, 0.009. The degree distribution diverges from a Poisson (which is the expected behavior in ER-like graphs) as we incorporate geographic features in the network, and the effect is especially significant when long edges are strongly penalized.</p> Full article ">Figure 6
<p>(<b>Florence</b>) Distribution of the size of the connected components other than the giant for friendship graphs with <math display="inline"><semantics> <mrow> <mi>μ</mi> <mo>=</mo> <mn>5</mn> </mrow> </semantics></math> and <math display="inline"><semantics> <mrow> <mi>μ</mi> <mo>=</mo> <mn>10</mn> </mrow> </semantics></math>. The lines show the mean, whereas the error bars (with “caps”) represent the 95% confidence interval computed over each set of runs. All components other than the giant are, in general, very small, but the use of real data does favor the occurrence of a higher number of connected components.</p> Full article ">Figure 7
<p>(<b>Florence</b>) Overview of the urban social graph. Each plot shows the average with confidence interval for 10 independent runs with the same configuration. The plots are discussed in <a href="#sec4dot2-futureinternet-13-00108" class="html-sec">Section 4.2</a> and <a href="#sec4dot3-futureinternet-13-00108" class="html-sec">Section 4.3</a>. (<b>a</b>) Size of the largest 50 connected components. (<b>b</b>) Size of the largest 50 clusters (Louvain). (<b>c</b>) Degree distribution with Lognormal fit. (<b>d</b>) Geographical distance between adjacent vertices. (<b>e</b>) Mean and max intra-cluster geographical distances. (<b>f</b>) Mean distance to all other vertices vs. Degree.</p> Full article ">Figure 8
<p>(<b>Florence</b>) Modularity of the obtained clustering structure—distribution for 10 independent runs per configuration. The quality of the obtained partition increases for small <math display="inline"><semantics> <mi>μ</mi> </semantics></math> and large <math display="inline"><semantics> <mi>β</mi> </semantics></math>, whereas it is not influenced by <math display="inline"><semantics> <msub> <mi>f</mi> <mi>u</mi> </msub> </semantics></math>.</p> Full article ">Figure 9
<p>(<b>Florence</b>) Adjacency matrix of the social graph with nodes (people) ordered by age-group. In both cases, <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math>, <span class="html-italic">μ</span> = 10 and <math display="inline"><semantics> <mrow> <mi>D</mi> <mo stretchy="false">(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo stretchy="false">)</mo> <mo>=</mo> <mi>d</mi> <msup> <mrow> <mo>(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo>)</mo> </mrow> <mrow> <mo>−</mo> <mi>β</mi> </mrow> </msup> </mrow> </semantics></math>, but β varies between the two figures. The assortativity by age is clearly visible in both cases.</p> Full article ">Figure 10
<p>(<b>Florence</b>) Heatmaps of the population and of the average degree per tile of the considered territory. The average degree is obtained for a graph with <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math>, <span class="html-italic">μ</span> = 10 and <math display="inline"><semantics> <mrow> <mi>D</mi> <mo stretchy="false">(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo stretchy="false">)</mo> <mo>=</mo> <mi>d</mi> <msup> <mrow> <mo>(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo>)</mo> </mrow> <mrow> <mo>−</mo> <mi>β</mi> </mrow> </msup> </mrow> </semantics></math>, for both <math display="inline"><semantics> <mrow> <mi>β</mi> <mo>∈</mo> <mo>{</mo> <mn>0.5</mn> <mo>,</mo> <mn>2</mn> <mo>}</mo> </mrow> </semantics></math>. With <math display="inline"><semantics> <mrow> <mi>β</mi> <mo>=</mo> <mn>0.5</mn> </mrow> </semantics></math> (<b>b</b>), the mean degree is weakly correlated with the population density, contrarily to what happens when <math display="inline"><semantics> <mrow> <mi>β</mi> <mo>=</mo> <mn>2</mn> </mrow> </semantics></math> (<b>c</b>).</p> Full article ">Figure 11
<p>(<b>Florence</b>) Impact of switching from <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>≡</mo> <mn>1</mn> </mrow> </semantics></math> to <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math> on the degree distribution of each tile. In both cases, <span class="html-italic">μ</span> = 10 and <math display="inline"><semantics> <mrow> <mi>D</mi> <mo stretchy="false">(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo stretchy="false">)</mo> <mo>=</mo> <mi>d</mi> <msup> <mrow> <mo>(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo>)</mo> </mrow> <mrow> <mo>−</mo> <mn>2</mn> </mrow> </msup> </mrow> </semantics></math>. <span class="html-italic">f<sub>u</sub></span> has no impact on the mean degree of each tile (<b>a</b>), but setting <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math> yields individuals with very different sociability inside each tile (b). (<b>a</b>) Difference of the <span class="html-italic">mean</span> degree of each tile between the configuration with <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math> and the configuration with <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>≡</mo> <mn>1</mn> </mrow> </semantics></math> (<b>b</b>) Difference of the <span class="html-italic">maximum</span> degree of each tile between the configuration with <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math> and the configuration with <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>≡</mo> <mn>1</mn> </mrow> </semantics></math>.</p> Full article ">Figure A1
<p>A graphical representation of the Territory of <b>Viterbo</b> (<b>a</b>) and of the population of <b>Florence</b> (<b>b</b>). (<b>a</b>) Random sample of 1000 people on the territory of Viterbo. (<b>b</b>) Population density for the city of Florence.</p> Full article ">Figure A2
<p>(<b>Florence</b>) Number of people not assigned to any household in different simulations, as the size of the population, the side of each tile and the noise added to the age distribution vary. (<b>a</b>) Number of unassigned people vs. population size. (<b>b</b>) Number of unassigned people vs. tile size (Florence).</p> Full article ">Figure A3
<p>(<b>Florence</b>) Distribution of household types and number of children obtained with our heuristic, with and without noise on the input data, compared with ISTAT data.</p> Full article ">Figure A4
<p>(<b>Viterbo</b>) Average degree of the individuals of each age group, in the whole friendship graph <math display="inline"><semantics> <msub> <mi>G</mi> <mi>F</mi> </msub> </semantics></math> and with their peers, under different configurations all with <math display="inline"><semantics> <mrow> <mi>μ</mi> <mo>=</mo> <mn>5</mn> </mrow> </semantics></math>.</p> Full article ">Figure A5
<p>(<b>Viterbo</b>) Percentage of edges between age groups in a configuration with data-driven population and age-based mixing, for <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math>, <span class="html-italic">μ</span> = 5 and <math display="inline"><semantics> <mrow> <mi>D</mi> <mo stretchy="false">(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo stretchy="false">)</mo> <mo>=</mo> <mi>d</mi> <msup> <mrow> <mo>(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo>)</mo> </mrow> <mrow> <mo>−</mo> <mn>2</mn> </mrow> </msup> </mrow> </semantics></math>.</p> Full article ">Figure A6
<p>(<b>Viterbo</b>) Degree distributions of friendship graphs with <math display="inline"><semantics> <mrow> <mi>μ</mi> <mo>=</mo> <mn>5</mn> </mrow> </semantics></math> and corresponding Poisson distributions with <math display="inline"><semantics> <mrow> <mi>λ</mi> <mo>=</mo> <mn>5</mn> </mrow> </semantics></math>. The data distribution lines show the mean, whereas the shaded areas around the lines represent the 95% confidence interval. The Kullback–Leibler divergence between the Poisson and the empirical distribution are (top to bottom, left to right): 8.4e-07, 7.2e-07, 6.03e-05, 0.001, 0.001, 0.03.</p> Full article ">Figure A7
<p>(<b>Viterbo</b>) Degree distributions of friendship graphs with <math display="inline"><semantics> <mrow> <mi>μ</mi> <mo>=</mo> <mn>10</mn> </mrow> </semantics></math> and corresponding Poisson distributions with <math display="inline"><semantics> <mrow> <mi>λ</mi> <mo>=</mo> <mn>10</mn> </mrow> </semantics></math>. The data distribution lines show the mean, whereas the shaded areas around the lines represent the 95% confidence interval. The Kullback–Leibler divergence between the Poisson and the empirical distribution are (left to right, top to bottom): 1.41e-06, 1.003e-06, 0.0002, 0.003, 0.003, 0.05.</p> Full article ">Figure A8
<p>(<b>Viterbo</b>) Distribution of the size of the connected components other than the giant for friendship graphs with <math display="inline"><semantics> <mrow> <mi>μ</mi> <mo>=</mo> <mn>5</mn> </mrow> </semantics></math> and <math display="inline"><semantics> <mrow> <mi>μ</mi> <mo>=</mo> <mn>10</mn> </mrow> </semantics></math>. The lines show the mean, whereas the error bars (with “caps”) represent the 95% confidence interval computed over each set of runs.</p> Full article ">Figure A9
<p>(<b>Viterbo</b>) Overview of the urban social graph for the city of Viterbo. Each plot shows the average with confidence interval for 10 independent runs with the same configuration. (<b>a</b>) Size of the largest 50 connected components. (<b>b</b>) Size of the largest 50 clusters (Louvain). (<b>c</b>) Degree distribution with Lognormal fit. (<b>d</b>) Geographical distance between adjacent vertices. (<b>e</b>) Mean and max intra-cluster geographical distances. (<b>f</b>) Mean distance to all other vertices vs. Degree.</p> Full article ">Figure A10
<p>(<b>Viterbo</b>) Modularity of the obtained clustering structure—distribution for 10 independent runs per configuration.</p> Full article ">Figure A11
<p>(<b>Viterbo</b>) Adjacency matrix of the social graph with nodes (people) ordered by agegroup. In both cases, <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math>, <span class="html-italic">μ</span> = 10 and <math display="inline"><semantics> <mrow> <mi>D</mi> <mo stretchy="false">(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo stretchy="false">)</mo> <mo>=</mo> <mi>d</mi> <msup> <mrow> <mo>(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo>)</mo> </mrow> <mrow> <mo>−</mo> <mi>β</mi> </mrow> </msup> </mrow> </semantics></math>, but β varies between the two figures.</p> Full article ">Figure A12
<p>(<b>Viterbo</b>) Heatmaps of the population and of the average degree per tile of the considered territory. The average degree is obtained for a graph with <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math>, <span class="html-italic">μ</span> = 10 and <math display="inline"><semantics> <mrow> <mi>D</mi> <mo stretchy="false">(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo stretchy="false">)</mo> <mo>=</mo> <mi>d</mi> <msup> <mrow> <mo>(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo>)</mo> </mrow> <mrow> <mo>−</mo> <mi>β</mi> </mrow> </msup> </mrow> </semantics></math>, for both <math display="inline"><semantics> <mrow> <mi>β</mi> <mo>∈</mo> <mo>{</mo> <mn>0.5</mn> <mo>,</mo> <mn>2</mn> <mo>}</mo> </mrow> </semantics></math>.</p> Full article ">Figure A13
<p>(<b>Viterbo</b>) Impact of switching from <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>≡</mo> <mn>1</mn> </mrow> </semantics></math> to <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math> on the degree distribution of each tile. In both cases, <span class="html-italic">μ</span> = 10 and <math display="inline"><semantics> <mrow> <mi>D</mi> <mo stretchy="false">(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo stretchy="false">)</mo> <mo>=</mo> <mi>d</mi> <msup> <mrow> <mo>(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo>)</mo> </mrow> <mrow> <mo>−</mo> <mn>2</mn> </mrow> </msup> </mrow> </semantics></math>. (<b>a</b>) Difference of the <span class="html-italic">mean</span> degree of each tile between the configuration with <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math> and the configuration with <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>≡</mo> <mn>1</mn> </mrow> </semantics></math> (<b>b</b>) Difference of the <span class="html-italic">maximum</span> degree of each tile between the configuration with <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math> and the configuration with <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>≡</mo> <mn>1</mn> </mrow> </semantics></math>.</p> Full article ">Figure A14
<p>(<b>Sabaudia</b>) Average degree of the individuals of each age group, in the whole friendship graph <math display="inline"><semantics> <msub> <mi>G</mi> <mi>F</mi> </msub> </semantics></math> and with their peers, under different configurations all with <math display="inline"><semantics> <mrow> <mi>μ</mi> <mo>=</mo> <mn>5</mn> </mrow> </semantics></math>.</p> Full article ">Figure A15
<p>(<b>Sabaudia</b>) Percentage of edges between age groups in a configuration with data-driven population and age-based mixing, for <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math>, <span class="html-italic">μ</span> = 5 and <math display="inline"><semantics> <mrow> <mi>D</mi> <mo stretchy="false">(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo stretchy="false">)</mo> <mo>=</mo> <mi>d</mi> <msup> <mrow> <mo>(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo>)</mo> </mrow> <mrow> <mo>−</mo> <mn>2</mn> </mrow> </msup> </mrow> </semantics></math>.</p> Full article ">Figure A16
<p>(<b>Sabaudia</b>) Degree distributions of friendship graphs with <math display="inline"><semantics> <mrow> <mi>μ</mi> <mo>=</mo> <mn>5</mn> </mrow> </semantics></math> and corresponding Poisson distributions with <math display="inline"><semantics> <mrow> <mi>λ</mi> <mo>=</mo> <mn>5</mn> </mrow> </semantics></math>. The data distribution lines show the mean, whereas the shaded areas around the lines represent the 95% confidence interval. The Kullback–Leibler divergence between the Poisson and the empirical distribution are (left to right, top to bottom): 5.3e-06, 5.8e-06, 1.2e-05, 0.001, 0.0001, 0.05.</p> Full article ">Figure A17
<p>(<b>Sabaudia</b>) Degree distributions of friendship graphs with <math display="inline"><semantics> <mrow> <mi>μ</mi> <mo>=</mo> <mn>5</mn> </mrow> </semantics></math> and corresponding Poisson distributions with <math display="inline"><semantics> <mrow> <mi>λ</mi> <mo>=</mo> <mn>5</mn> </mrow> </semantics></math>. The data distribution lines show the mean, whereas the shaded areas around the lines represent the 95% confidence interval. The Kullback–Leibler divergence between the Poisson and the empirical distribution are (left to right, top to bottom): 2.23e-06, 3.01e-06, 3.4e-05, 0.003, 0.0004, 0.08.</p> Full article ">Figure A18
<p>(<b>Sabaudia</b>) Distribution of the size of the connected components other than the giant for friendship graphs with <math display="inline"><semantics> <mrow> <mi>μ</mi> <mo>=</mo> <mn>5</mn> </mrow> </semantics></math> and <math display="inline"><semantics> <mrow> <mi>μ</mi> <mo>=</mo> <mn>10</mn> </mrow> </semantics></math>. The lines show the mean, whereas the error bars (with “caps”) represent the 95% confidence interval computed over each set of runs.</p> Full article ">Figure A19
<p>(<b>Sabaudia</b>) Overview of the urban social graph for the city of Sabaudia. Each plot shows the average with confidence interval for 10 independent runs with the same configuration. (<b>a</b>) Size of the largest 50 connected components. (<b>b</b>) Size of the largest 50 clusters (Louvain). (<b>c</b>) Degree distribution with Lognormal fit. (<b>d</b>) Geographical distance between adjacent vertices. (<b>e</b>) Mean and max intra-cluster geographical distances. (<b>f</b>) Mean distance to all other vertices vs. Degree.</p> Full article ">Figure A20
<p>(<b>Sabaudia</b>) Modularity of the obtained clustering structure—distribution for 10 independent runs per configuration.</p> Full article ">Figure A21
<p>(<b>Sabaudia</b>) Adjacency matrix of the social graph with nodes (people) ordered by agegroup. In both cases, <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math>, <span class="html-italic">μ</span> = 10 and <math display="inline"><semantics> <mrow> <mi>D</mi> <mo stretchy="false">(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo stretchy="false">)</mo> <mo>=</mo> <mi>d</mi> <msup> <mrow> <mo>(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo>)</mo> </mrow> <mrow> <mo>−</mo> <mi>β</mi> </mrow> </msup> </mrow> </semantics></math>, but β varies between the two figures.</p> Full article ">Figure A22
<p>(<b>Sabaudia</b>) Heatmaps of the population and of the average degree per tile of the considered territory. The average degree is obtained for a graph with <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math>, <span class="html-italic">μ</span> = 10 and <math display="inline"><semantics> <mrow> <mi>D</mi> <mo stretchy="false">(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo stretchy="false">)</mo> <mo>=</mo> <mi>d</mi> <msup> <mrow> <mo>(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo>)</mo> </mrow> <mrow> <mo>−</mo> <mi>β</mi> </mrow> </msup> </mrow> </semantics></math>, for both <math display="inline"><semantics> <mrow> <mi>β</mi> <mo>∈</mo> <mo>{</mo> <mn>0.5</mn> <mo>,</mo> <mn>2</mn> <mo>}</mo> </mrow> </semantics></math>.</p> Full article ">Figure A23
<p>(<b>Sabaudia</b>) Impact of switching from <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>≡</mo> <mn>1</mn> </mrow> </semantics></math> to <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math> on the degree distribution of each tile. In both cases, <span class="html-italic">μ</span> = 10 and <math display="inline"><semantics> <mrow> <mi>D</mi> <mo stretchy="false">(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo stretchy="false">)</mo> <mo>=</mo> <mi>d</mi> <msup> <mrow> <mo>(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo>)</mo> </mrow> <mrow> <mo>−</mo> <mn>2</mn> </mrow> </msup> </mrow> </semantics></math>. (<b>a</b>) Difference of the <span class="html-italic">mean</span> degree of each tile between the configuration with <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math> and the configuration with <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>≡</mo> <mn>1</mn> </mrow> </semantics></math> (<b>b</b>) Difference of the <span class="html-italic">maximum</span> degree of each tile between the configuration with <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>∼</mo> <mn>1</mn> <mo>+</mo> <mi mathvariant="script">LN</mi> <mrow> <mo>(</mo> <mo form="prefix">ln</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> <mo>,</mo> <mn>0.25</mn> <mo>)</mo> </mrow> </mrow> </semantics></math> and the configuration with <math display="inline"><semantics> <mrow> <msub> <mi>f</mi> <mi>u</mi> </msub> <mo>≡</mo> <mn>1</mn> </mrow> </semantics></math>.</p> Full article ">
15 pages, 596 KiB  
Article
Mutual Influence of Users Credibility and News Spreading in Online Social Networks
by Vincenza Carchiolo, Alessandro Longheu, Michele Malgeri, Giuseppe Mangioni and Marialaura Previti
Future Internet 2021, 13(5), 107; https://doi.org/10.3390/fi13050107 - 25 Apr 2021
Cited by 9 | Viewed by 3220
Abstract
A real-time news spreading is now available for everyone, especially thanks to Online Social Networks (OSNs) that easily endorse gate watching, so the collective intelligence and knowledge of dedicated communities are exploited to filter the news flow and to highlight and debate relevant [...] Read more.
A real-time news spreading is now available for everyone, especially thanks to Online Social Networks (OSNs) that easily endorse gate watching, so the collective intelligence and knowledge of dedicated communities are exploited to filter the news flow and to highlight and debate relevant topics. The main drawback is that the responsibility for judging the content and accuracy of information moves from editors and journalists to online information users, with the side effect of the potential growth of fake news. In such a scenario, trustworthiness about information providers cannot be overlooked anymore, rather it more and more helps in discerning real news from fakes. In this paper we evaluate how trustworthiness among OSN users influences the news spreading process. To this purpose, we consider the news spreading as a Susceptible-Infected-Recovered (SIR) process in OSN, adding the contribution credibility of users as a layer on top of OSN. Simulations with both fake and true news spreading on such a multiplex network show that the credibility improves the diffusion of real news while limiting the propagation of fakes. The proposed approach can also be extended to real social networks. Full article
(This article belongs to the Special Issue Digital and Social Media in the Disinformation Age)
►▼ Show Figures

Figure 1

Figure 1
<p>A two-layers representation of our model: online social network (OSN) and credibility network (CN) layers.</p> Full article ">Figure 2
<p>Direct credibility <math display="inline"><semantics> <msub> <mi>C</mi> <mrow> <mi>j</mi> <mi>i</mi> </mrow> </msub> </semantics></math> dynamics, x-axis shows the time and y-axis the node credibility.</p> Full article ">Figure 3
<p>Temporal evolution of the average number of infected nodes in different scenarios.</p> Full article ">Figure 4
<p>Incoming credibility strength distribution for cases B and D during the 1st, 50th, 100th and 150th news.</p> Full article ">
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
 
Displaying articles 1-34
Previous Issue
Next Issue
Back to TopTop