CN113315757A - Data decoding-free transmission method facing edge calculation - Google Patents
Data decoding-free transmission method facing edge calculation Download PDFInfo
- Publication number
- CN113315757A CN113315757A CN202110503683.2A CN202110503683A CN113315757A CN 113315757 A CN113315757 A CN 113315757A CN 202110503683 A CN202110503683 A CN 202110503683A CN 113315757 A CN113315757 A CN 113315757A
- Authority
- CN
- China
- Prior art keywords
- data
- edge
- matrix
- neighbor
- decoding
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L69/00—Network arrangements, protocols or services independent of the application payload and not provided for in the other groups of this subclass
- H04L69/04—Protocols for data compression, e.g. ROHC
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Computer Security & Cryptography (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
A decoding-free transmission method of compressed data facing to edge calculation collects data of different time slots on edge fixed network or mobile network at initial node, stores them in different arrays respectively, constructs neighbor weighted graph for data in each array, records a pair of data with adjacent relation as data pair, records the adjacent relation between corresponding characteristic attributes of data pair, constructs global measurement characteristic matrix, solves to obtain the shortest path of all data pairs in neighbor weighted graph, captures characteristic mapping embedding, maps data characteristic from high-dimensional space to low-dimensional space, obtains characteristic mapping result of all time slots, combines the result as the compressed expression result of characteristic mapping of next time slot edge network node, and combines and transmits to edge network node, the invention needs no data decoding, can analyze and process data accurately, the computing efficiency is greatly improved, computing resources are saved, and the transmission performance of mass data in cloud computing is improved.
Description
Technical Field
The invention belongs to the technical field of industrial automation and cloud computing, relates to data processing and transmission, and particularly relates to a compressed data decoding-free transmission method facing edge computing.
Background
With the development of the 5G technology, a great amount of information data is generated in the mobile communication network, because the information data are various in types and have high propagation speed, the network data processing consumes more time and computing resources, and one of the great advantages of edge computing is that the data can be processed and stored more quickly. Edge computing has become a classic paradigm to increase the computing power of a mobile device and reduce its energy consumption, and in the edge computing paradigm, a mobile device migrates cloud-centric computing tasks onto nearby edge servers to alleviate cloud-centric loads while helping to reduce the energy consumption of the device and the execution latency of the tasks. Large-scale data information is transmitted from a cloud center to an edge network, and a data compression technology becomes an effective measure for solving the problem of transmission and processing of the current mass data. The compact data representation not only can improve the data transmission efficiency, but also can flexibly and efficiently process data. At present, a data compression algorithm for edge calculation firstly decodes compressed data and then analyzes and processes the data, however, network resource nodes in the edge calculation are limited, a large amount of network transmission delay is generated in the data decoding process, and efficient processing operation of the compressed data by an edge network is not facilitated.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention aims to provide a decoding-free transmission method for compressed data facing edge computing, which directly analyzes and classifies the compressed data without data decoding, greatly improves the computing efficiency, saves computing resources and improves the transmission performance of mass data in cloud computing on the premise of accurately analyzing and processing the data.
In order to achieve the purpose, the invention adopts the technical scheme that:
an edge-computation-oriented decoding-free transmission method for compressed data, comprising:
step 1, collecting data of different time slots on an edge fixed network or a mobile network at an initial node;
step 2, adopting a classifiable manifold learning data compression algorithm to compress the collected data, and comprising the following steps:
step 2.1, respectively storing the data of different time slots collected in the step 1 in different arrays;
step 2.2, constructing a neighbor weighted graph G for the data in each array, recording a pair of data with adjacent relation as a data pair, and recording the adjacent relation between the characteristic attributes corresponding to the data pair;
step 2.3, constructing a global measurement characteristic matrix, and solving the global measurement characteristic matrix by using a geodesic distance algorithm to obtain the shortest paths of all data pairs in the neighbor weighted graph G;
step 2.4, capturing feature mapping embedding, and mapping the data features from a high-dimensional space to a low-dimensional space;
step 2.5, obtaining the feature mapping results of all time slots, and combining the results to be used as the feature mapping compression expression result of the next time slot edge network node;
and step 3, transmitting the result combination to the edge network node.
In the step 1, data on the edge fixed network or the mobile network is collected according to the time slot, and the characteristic attribute of the data is recorded, and in the step 2.1, the data in the edge fixed network or the mobile network of different time slots are respectively recorded by using different arrays.
In the step 2.2, a neighbor weighted graph G is constructed by using an epsilon-neighbor method, and the method comprises the following steps:
given a threshold ε, if | | | xi-xj||2If epsilon is not more than epsilon, data xiAnd data xjIs a neighbor point, has a neighbor relation, and is marked as a data pair, wherein xiCorresponding to the ith characteristic attribute ai，xjCorresponding to the ith characteristic attribute ajX is to beiAnd xjConnected by an edge, and the weight L of the edge is set as xiAnd xjEuclidean distance of dx(i, j) which represents aiAnd ajIn a manner such that, among other things,||||2representing the euclidean norm.
In step 2.3, the global metric feature matrix is expressed as: dG＝(dG(i,j))n*nThe geodesic distance algorithm solving formula is as follows:
wherein D isGN x n matrix composed of shortest paths of all data pairs in neighbor weighted graph G, dG(i, j) represents xiAnd xjGeodesic distance between, N represents a data set, dG(i, k) represents data xiAnd data xkGeodesic distance between, dG(k, j) represents data xkAnd data xjGeodesic distance between them.
In the step 2.4, order
Find matrix τ (D)G) All the eigenvalues and eigenvectors are sorted from large to small according to the eigenvalues, and the first d eigenvalues lambda are selected1,λ2…λdUsing λ1,λ2…λdCorresponding feature vector u1,u2…udThe composition matrix U ═ U1,u2…ud]Then, the final feature mapping embedding result is:
where Y represents the low-dimensional eigenmap embedding result, i.e., eigenmap result, H is the center matrix, which is summed with DGIs a unit matrix of the same order, S is a square matrix of the geodesic distance,
τ denotes a matrix transform operator, and C denotes an adjacent weighted graph G.
Compared with the prior art, the invention has the beneficial effects that:
the data compression method adopted by the invention omits the decoding operation of the compressed data in the data processing process by utilizing the characteristic mapping compression algorithm, and finally obviously optimizes the data transmission and processing time and saves the calculation and communication resources on the basis of improving the classification accuracy.
Drawings
FIG. 1 is a schematic flow diagram of the present invention.
Fig. 2 is a detailed flow chart of an embodiment of the present invention.
Detailed Description
The embodiments of the present invention will be described in detail below with reference to the drawings and examples.
As shown in fig. 1, the present invention is a decoding-free transmission method for compressed data facing edge calculation, which mainly includes the following steps:
step 1, at an initial node, collecting data of time slots 1-t on an edge fixed network or a mobile network, and recording the characteristic attribute and the corresponding data value of the data.
Step 2, in order to reduce the transmission data scale and optimize the transmission efficiency, a classifiable manifold learning data compression algorithm is adopted at an initial node to perform feature mapping on data homomorphism, so that data compression is effectively realized on the premise of ensuring the classifiable data, and the premise is provided for efficiently acquiring data by an edge node, wherein the method comprises the following steps:
and 2.1, respectively storing the data of the time slots 1 to t collected in the step 1 in different arrays, namely storing the data of the time slot 1 in the array 1, storing the data of the time slot 2 in the arrays 2 and … …, and so on.
And 2.2, constructing a neighbor weighted graph G for the data in each array, recording a pair of data with adjacent relation as a data pair, and recording the adjacent relation between the characteristic attributes corresponding to the data pair.
The neighbor weighted graph G is constructed by an epsilon-neighbor method, and the method comprises the following steps:
in some array, given a threshold ε, if | | xi-xj||2If epsilon is not more than epsilon, data xiAnd data xjIs a neighbor point, has a neighbor relation, and is marked as a data pair, wherein xiThe corresponding characteristic attribute is ai，xjThe corresponding characteristic attribute is ajX is to beiAnd xjConnected by an edge, and the weight L of the edge is set as xiAnd xjEuclidean distance of dx(i, j) which represents aiAnd ajHave an adjacent relationship therebetween, wherein | | | | non-phosphor2Representing the euclidean norm.
Step 2.3, constructing a global metric feature matrix, expressed as:
DG＝(dG(i,j))n*n
and (3) solving the global measurement characteristic matrix by using a geodesic distance algorithm to obtain the shortest paths of all data pairs in the neighbor weighted graph G, wherein the solving formula is as follows:
wherein D isGN x n matrix composed of shortest paths of all data pairs in neighbor weighted graph G, dG(i, j) represents xiAnd xjGeodesic distance between, N represents a data set, dG(i, k) represents data xiAnd data xkGeodesic distance between, dG(k, j) represents data xkAnd data xjGeodesic distance between them.
And 2.4, capturing feature mapping embedding, and mapping the data features from a high-dimensional space to a low-dimensional space.
Specifically, let
Solving matrix tau (D) by knowledge of matrix analysisG) All the eigenvalues and eigenvectors are sorted from large to small according to the eigenvalues, and the first d eigenvalues lambda are selected1,λ2…λdUsing λ1,λ2…λdCorresponding feature vector u1,u2…udThe composition matrix U ═ U1,u2…ud]Then, the final feature mapping embedding result is:
where Y represents the low-dimensional eigenmap embedding result, i.e., eigenmap result, H is the center matrix, which is summed with DGIs a unit matrix of the same order, S is a square matrix of the geodesic distance,
τ denotes a matrix transform operator, and C denotes an adjacent weighted graph G.
And 2.5, sequentially calculating and obtaining feature mapping results Y1-Yt of the time slots 1-t according to the method, taking (Y1, Y2, … … and Yt) as a feature mapping compression representation result of the edge network node of the time slot t +1, and transmitting the result to the edge network node.
The invention can adopt an Options-based data processing method to perform time domain abstraction at the central node in the transmission process, and uses the Options to output data compression classification results of different time slots. Therefore, on the premise of data compression, the classification accuracy and efficiency are effectively improved, and computing and communication resources are saved.
In the present invention, Options are defined on MDP (Markov Decision Process) and are used
A triple is represented by a single triple, and,
representing the initial state set, pi representing the policy, and beta representing the terminal state set. ε (o, S, t) represents the probability that the options end up in state S after the compressed data classification model performs k steps, for each state S ∈ S, p (S, k) represents the probability that the options end up in state S,after terminating one state, the next state is entered. The equation for the state transition function for the state s with respect to option is as follows:
wherein gamma represents the threshold value of the threshold value,
indicating the next state. The method is used for describing the classification output results of the compressed data classification model on different time scales.
Referring to fig. 2, in one embodiment of the present invention, two data sets, respectively data of a rossford national forest in northern colorado and athletic data of a plurality of healthy elderly people, are used, as shown in tables 1 and 2, respectively. The dataset is a Rossfu national forest in northern Colorado, with 4 wildland types, including 6 types of forest coverage, each of which will be determined by 12 data features. And the data set II records the data change of daily activities of the old through a method of installing sensors on coats worn by a plurality of old people. The sensor recording contents are respectively as follows: time, equalized reading G on the positive axis, equalized reading G on the vertical axis, Sensor ID on the absolute reading, Signal string, Phase, and Frequency. The collected attribute features are then deposited in an array. The twelve characteristic attributes of the forest comprise factors of Elevation, Aspect, Slope, Hillshade _9am, Wilderness _ Area, Soil _ Type and the like. And for the first data set, setting a plurality of arrays, wherein the size of each array is 12, and the arrays are used for storing 12 characteristic attributes of the forest to record 12 characteristic attributes of different time slots. For data set two, the array size is set to 8, for storing the attributes of 8 elderly activities.
TABLE 1 four wilderness region information table
Table 2 data set information table
| Datasets 1 | Size | Datasets 2 | Size |
| DS_10 | 7000 | DS_10 | 4000 |
| DS_20 | 14000 | DS_20 | 8000 |
| DS_30 | 21000 | DS_30 | 12000 |
| DS_40 | 28000 | DS_40 | 16000 |
| DS_50 | 35000 | DS_50 | 20000 |
After the feature mapping compression representation result is obtained by the method, a classification model of compressed data is constructed by using a BP neural network and an SVM respectively, the data of the two compressed data sets are classified, the classified result is compared with the original class, and the classification performance of the compressed data is verified.
To verify the effectiveness of the present invention, two comparison schemes were set up: the contrast scheme is data compression realized based on a self-encoder, and the contrast scheme does not perform any compression processing on original input data. Wherein the scheme used by the invention is represented by the ISOMAP-C @ BP and ISOMAP-C @ SVM symbols. According to the comparison scheme, a BP neural network and an SVM are combined to build a classification model respectively, wherein the CAE @ BP mark utilizes a BP neural network algorithm to build a scheme of a classification model facing a compression result, and the CAE @ SVM mark SVM is used to build a scheme of a classification model facing the compression result; and in the second comparison scheme, original input data are directly and respectively classified by a BP classifier and an SVM classifier and are marked by Basic @ BP and Basic @ SVM. The evaluation indexes adopted are accuracy and efficiency.
Table 3 table of accuracy comparison under data set 1
Table 4 table of accuracy comparison under data set 2
Table 5 run time comparison table under data set 1
Table 6 specific run time comparison table under dataset 2
The invention can remarkably optimize the data transmission or processing time by the proposed classifiable data compression mechanism on the premise of improving the classification accuracy by 2.2%.
While the invention has been described in detail with reference to specific embodiments thereof, it will be understood that the invention is not limited to the details of construction and the embodiments set forth herein. For a person skilled in the art to which the invention pertains, several simple deductions or substitutions may be made without departing from the spirit of the invention and the scope of protection defined by the claims, which shall be regarded as belonging to the scope of protection of the invention.
Claims (5)
1. An edge-computation-oriented decoding-free transmission method for compressed data, comprising:
step 1, collecting data of different time slots on an edge fixed network or a mobile network at an initial node;
step 2, adopting a classifiable manifold learning data compression algorithm to compress the collected data, and comprising the following steps:
step 2.1, respectively storing the data of different time slots collected in the step 1 in different arrays;
step 2.2, constructing a neighbor weighted graph G for the data in each array, recording a pair of data with adjacent relation as a data pair, and recording the adjacent relation between the characteristic attributes corresponding to the data pair;
step 2.3, constructing a global measurement characteristic matrix, and solving the global measurement characteristic matrix by using a geodesic distance algorithm to obtain the shortest paths of all data pairs in the neighbor weighted graph G;
step 2.4, capturing feature mapping embedding, and mapping the data features from a high-dimensional space to a low-dimensional space;
step 2.5, obtaining the feature mapping results of all time slots, and combining the results to be used as the feature mapping compression expression result of the next time slot edge network node;
and step 3, transmitting the result combination to the edge network node.
2. The decoding-free transmission method of compressed data for edge-oriented computing according to claim 1 or 2, wherein in the step 1, data on an edge fixed network or a mobile network is collected according to a time slot, and characteristic attributes of the data are recorded, and in the step 2.1, data in edge fixed networks or mobile networks with different time slots are respectively recorded by using different arrays.
3. The decoding-free transmission method of compressed data for edge-oriented computation according to claim 2, wherein in the step 2.2, a neighbor weighting graph G is constructed by using an epsilon-neighbor method, and the method is as follows:
given a threshold ε, if | | | xi-xj||2If epsilon is not more than epsilon, data xiAnd data xjIs a neighbor point, has a neighbor relation, and is marked as a data pair, wherein xiThe corresponding characteristic attribute is ai，xjThe corresponding characteristic attribute is ajX is to beiAnd xjConnected by an edge, and the weight L of the edge is set as xiAnd xjEuclidean distance of dx(i, j) which represents aiAnd ajA relation between | | | non-calculation2Representing the euclidean norm.
4. The decoding-free transmission method of compressed data for edge-oriented computation according to claim 3, wherein in the step 2.3, the global metric feature matrix is expressed as: dG＝(dG(i，j))n*nThe geodesic distance algorithm solving formula is as follows:
wherein D isGN x n matrix composed of shortest paths of all data pairs in neighbor weighted graph G, dG(i, j) represents xiAnd xjGeodesic distance between, N represents a data set, dG(i, k) represents data xiAnd data xkGeodesic distance between, dG(k, j) represents data xkAnd data xjGeodesic distance between them.
5. The decoding-free transmission method of compressed data for edge-oriented computation of claim 4, wherein in the step 2.4, the step of compressing the compressed data for edge-oriented computation comprises
Find matrix τ (D)G) All the eigenvalues and eigenvectors are sorted from large to small according to the eigenvalues, and the first d eigenvalues lambda are selected1，λ2...λdUsing λ1，λ2...λdCorresponding feature vector u1，u2...udThe composition matrix U ═ U1，u2...ud]Then, the final feature mapping embedding result is:
where Y represents the low-dimensional eigenmap embedding result, i.e., eigenmap result, H is the center matrix, which is summed with DGIs a unit matrix of the same order, S is a square matrix of the geodesic distance,
τ denotes a matrix transform operator, and C denotes an adjacent weighted graph G.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110503683.2A CN113315757B (en) | 2021-05-10 | 2021-05-10 | Data decoding-free transmission method facing edge calculation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110503683.2A CN113315757B (en) | 2021-05-10 | 2021-05-10 | Data decoding-free transmission method facing edge calculation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113315757A true CN113315757A (en) | 2021-08-27 |
| CN113315757B CN113315757B (en) | 2022-04-08 |
Family
ID=77371759
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110503683.2A Active CN113315757B (en) | 2021-05-10 | 2021-05-10 | Data decoding-free transmission method facing edge calculation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113315757B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7218784B1 (en) * | 2000-05-01 | 2007-05-15 | Xerox Corporation | Method and apparatus for controlling image quality and compression ratios |
| US20100128949A1 (en) * | 2008-11-26 | 2010-05-27 | Samplify Systems, Inc. | Compression and storage of projection data in a computed tomography system |
| CN102184349A (en) * | 2011-04-29 | 2011-09-14 | 河海大学 | System and method for clustering gene expression data based on manifold learning |
| CN108901046A (en) * | 2018-06-14 | 2018-11-27 | 北京大学 | Cotasking unloading algorithm and system design scheme towards mobile edge calculations |
| CN110490231A (en) * | 2019-07-17 | 2019-11-22 | 哈尔滨工程大学 | A kind of Netflow Method of Data with Adding Windows for thering is supervision to differentiate manifold learning |
| CN110852962A (en) * | 2019-10-29 | 2020-02-28 | 南京邮电大学 | Dual-mapping learning compressed face image restoration method based on regression tree classification |
| CN110852304A (en) * | 2019-11-22 | 2020-02-28 | 重庆大学 | Hyperspectral data processing method based on deep learning method |
| CN111479286A (en) * | 2020-02-26 | 2020-07-31 | 国网河南省电力公司电力科学研究院 | Data processing method for reducing communication flow of edge computing system |
| CN111679904A (en) * | 2020-03-27 | 2020-09-18 | 北京世纪互联宽带数据中心有限公司 | Task scheduling method and device based on edge computing network |
-
2021
- 2021-05-10 CN CN202110503683.2A patent/CN113315757B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7218784B1 (en) * | 2000-05-01 | 2007-05-15 | Xerox Corporation | Method and apparatus for controlling image quality and compression ratios |
| US20100128949A1 (en) * | 2008-11-26 | 2010-05-27 | Samplify Systems, Inc. | Compression and storage of projection data in a computed tomography system |
| CN102184349A (en) * | 2011-04-29 | 2011-09-14 | 河海大学 | System and method for clustering gene expression data based on manifold learning |
| CN108901046A (en) * | 2018-06-14 | 2018-11-27 | 北京大学 | Cotasking unloading algorithm and system design scheme towards mobile edge calculations |
| CN110490231A (en) * | 2019-07-17 | 2019-11-22 | 哈尔滨工程大学 | A kind of Netflow Method of Data with Adding Windows for thering is supervision to differentiate manifold learning |
| CN110852962A (en) * | 2019-10-29 | 2020-02-28 | 南京邮电大学 | Dual-mapping learning compressed face image restoration method based on regression tree classification |
| CN110852304A (en) * | 2019-11-22 | 2020-02-28 | 重庆大学 | Hyperspectral data processing method based on deep learning method |
| CN111479286A (en) * | 2020-02-26 | 2020-07-31 | 国网河南省电力公司电力科学研究院 | Data processing method for reducing communication flow of edge computing system |
| CN111679904A (en) * | 2020-03-27 | 2020-09-18 | 北京世纪互联宽带数据中心有限公司 | Task scheduling method and device based on edge computing network |
Non-Patent Citations (2)
| Title |
|---|
| XIAOLIN ZHANG .ETL: "Analysis of the landscape pattern in eastern edge of Pearl River Delta based on the Object-oriented classification method: A case study of Huicheng District, Huizhou City", 《2011 INTERNATIONAL CONFERENCE ON REMOTE SENSING, ENVIRONMENT AND TRANSPORTATION ENGINEERING》 * |
| 陈晓江 等: "基于边缘计算技术的智能集中器多元负荷管理设计", 《电测与仪表》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113315757B (en) | 2022-04-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Gilbert et al. | Trust based data prediction, aggregation and reconstruction using compressed sensing for clustered wireless sensor networks | |
| EP3754906B1 (en) | Method and system for performing automated root cause analysis of anomaly events in high-dimensional sensor data | |
| Li et al. | A comparative analysis of evolutionary and memetic algorithms for community detection from signed social networks | |
| CN110138595A (en) | Time link prediction technique, device, equipment and the medium of dynamic weighting network | |
| Cheng et al. | An intelligent scheme for big data recovery in Internet of Things based on multi-attribute assistance and extremely randomized trees | |
| CN116842459B (en) | Electric energy metering fault diagnosis method and diagnosis terminal based on small sample learning | |
| Yang et al. | Joltik: enabling energy-efficient" future-proof" analytics on low-power wide-area networks | |
| Tang et al. | Reputation-aware data fusion and malicious participant detection in mobile crowdsensing | |
| Sekar et al. | Energy efficient data gathering using spatio-temporal compressive sensing for WSNs | |
| Jiang et al. | A graph-based approach for missing sensor data imputation | |
| Yu et al. | Rainfall time series forecasting based on Modular RBF Neural Network model coupled with SSA and PLS | |
| CN113315757B (en) | Data decoding-free transmission method facing edge calculation | |
| Zhao et al. | A novel optimization method for WSN based on mixed matrix decomposition of NMF and 2-SVD-QR | |
| CN107071858A (en) | A kind of subdivision remote sensing image method for parallel processing under Hadoop | |
| Munshi et al. | Comparisons among Bat algorithms with various objective functions on grouping photovoltaic power patterns | |
| Srinivas et al. | Detection of energy efficient sensor node in ehwsn | |
| Sandhu et al. | A novel framework for reliable network prediction of small scale wireless sensor networks (SSWSNs) | |
| CN114492569B (en) | Typhoon path classification method based on width learning system | |
| Huang et al. | A prediction scheme with genetic neural network and Isomap algorithm for tropical cyclone intensity change over western North Pacific | |
| CN115034839A (en) | Office area state detection method and device, storage medium and electronic equipment | |
| Putri et al. | Data prediction-based virtual sensor in digital twin scenario using deep learning approach | |
| Gupta et al. | An energy-efficient hierarchical data fusion approach in IoT | |
| Behzad et al. | Performance optimization in IoT-based next-generation wireless sensor networks | |
| Zordan et al. | Automatic rate-distortion classification for the IoT: Towards signal-adaptive network protocols | |
| CN117834455B (en) | Electric power Internet of things data transmission simulation method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |