[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.
 
 
Sign in to use this feature.

Years

Between: -

Subjects

remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline

Journals

remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline

Article Types

Countries / Regions

remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline

Search Results (1,368)

Search Parameters:
Keywords = fuzzy learning

Order results
Result details
Results per page
Select all
Export citation of selected articles as:
16 pages, 6672 KiB  
Article
Application of Fuzzy Adaptive Impedance Control Based on Backstepping Method for PAM Elbow Exoskeleton in Rehabilitation
by Zhirui Zhao, Xinyu Hou, Dexing Shan, Hongjun Liu, Hongshuai Liu and Lina Hao
Polymers 2024, 16(24), 3533; https://doi.org/10.3390/polym16243533 - 18 Dec 2024
Viewed by 218
Abstract
In this study, a fuzzy adaptive impedance control method integrating the backstepping control for the PAM elbow exoskeleton was developed to facilitate robot-assisted rehabilitation tasks. The proposed method uses fuzzy logic to adjust impedance parameters, thereby optimizing user adaptability and reducing interactive torque, [...] Read more.
In this study, a fuzzy adaptive impedance control method integrating the backstepping control for the PAM elbow exoskeleton was developed to facilitate robot-assisted rehabilitation tasks. The proposed method uses fuzzy logic to adjust impedance parameters, thereby optimizing user adaptability and reducing interactive torque, which are major limitations of traditional impedance control methods. Furthermore, a repetitive learning algorithm and an adaptive control strategy were incorporated to improve the performance of position accuracy, addressing the time-varying uncertainties and nonlinear disturbances inherent in the exoskeleton. The stability of the proposed controller was tested, and then corresponding simulations and an elbow flexion and extension rehabilitation experiment were performed. The results showed that, with the proposed method, the root mean square of the tracking error was 0.032 rad (i.e., 21.95% less than that of the PID method), and the steady-state interactive torque was 1.917 N·m (i.e., 46.49% less than that of the traditional impedance control). These values exceeded those of the existing methods and supported the potential application of the proposed method for other soft actuators and robots. Full article
(This article belongs to the Special Issue Advancing Soft Robotics with Polymers)
Show Figures

Figure 1

Figure 1
<p>Experimental results: interactive torque ((<b>a</b>) prototype of PAM elbow exoskeleton, (1) strap, (2) linkage, (3) PAM, (4) connector, and (5) handle); (<b>b</b>) design drawing.</p>
Full article ">Figure 2
<p>Two-layer architecture control structure: impedance control method (higher layer) and position control (lower layer).</p>
Full article ">Figure 3
<p>Two-layer architecture control structure: variable impedance control method based on fuzzy rules (higher layer) with backstepping repetitive learning control (lower layer).</p>
Full article ">Figure 4
<p>Membership function. (<b>a</b>) <math display="inline"><semantics> <mrow> <mo>Δ</mo> <mi>q</mi> </mrow> </semantics></math>, (<b>b</b>) <math display="inline"><semantics> <mrow> <mo>Δ</mo> <mover accent="true"> <mi>q</mi> <mo>˙</mo> </mover> </mrow> </semantics></math>, (<b>c</b>) <math display="inline"><semantics> <mrow> <mo>Δ</mo> <msub> <mi>τ</mi> <mi>e</mi> </msub> </mrow> </semantics></math>, and (<b>d</b>) <math display="inline"><semantics> <mrow> <mo>Δ</mo> <mi>K</mi> </mrow> </semantics></math>, <math display="inline"><semantics> <mrow> <mo>Δ</mo> <mi>D</mi> </mrow> </semantics></math>.</p>
Full article ">Figure 5
<p>Backstepping repetitive learning control (lower layer).</p>
Full article ">Figure 6
<p>The results of the simulation ((<b>a</b>) tracking trajectory and (<b>b</b>) position error).</p>
Full article ">Figure 7
<p>The results of the simulation ((<b>a</b>) damping and (<b>b</b>) stiffness).</p>
Full article ">Figure 8
<p>Experimental setups. (1) Air source. (2) Air filter regulator. (3) Valves. (4) PAM. (5) PAM elbow exoskeleton. (6) IMU. (7) Force sensor. (8) PC and screen. (9) Arduino mega 2560. (10) DC modular. (11) PWM modular. The black arrow indicates the power supply, and the dashed arrow indicates the implement.</p>
Full article ">Figure 9
<p>Rehabilitation experiment.</p>
Full article ">Figure 10
<p>Experimental results. Parameters: (<b>a</b>) joint position; (<b>b</b>) position error (between human intention and actual position).</p>
Full article ">Figure 11
<p>Experimental results. Parameters: (<b>a</b>) stiffness; (<b>b</b>) damping.</p>
Full article ">Figure 12
<p>Experimental results: interactive torque (<b>a</b>) traditional method; (<b>b</b>) proposed method.</p>
Full article ">
21 pages, 8572 KiB  
Article
The Measurement of Metal Mineral Particle Size Under the Microscope Based on Gaussian Pyramids and Directional Maximum Intercept
by Chaoxi Luo, Feng Xie, Bo Li, Xiangwen Lv, Meiguang Jiang, Jing Zhang, Sheng Jian, Fang Yang and Yong Wang
Minerals 2024, 14(12), 1284; https://doi.org/10.3390/min14121284 - 17 Dec 2024
Viewed by 296
Abstract
With the development of mineral resources, minerals are becoming increasingly difficult to process. In order to utilize these resources more effectively, in-depth research into process mineralogy has become increasingly important in the field of mineralogy, and particle size measurement under the microscope is [...] Read more.
With the development of mineral resources, minerals are becoming increasingly difficult to process. In order to utilize these resources more effectively, in-depth research into process mineralogy has become increasingly important in the field of mineralogy, and particle size measurement under the microscope is one of the critical aspects of process mineralogy. At present, the use of scanning electron microscopes and other equipment for measurement is very expensive, and manual measurement has problems such as poor accuracy and low efficiency. In addition, there is a lack of reference materials for the segmentation algorithm of mineral light images. This article proposes a Gaussian pyramid based on bilateral filtering combined with directional maximum intercept to measure mineral particle size under the microscope. In the experiments, different segmentation algorithms were studied, including Gaussian pyramid segmentation based on bilateral filtering, segmentation based on Fuzzy C-Means, and the rapidly developing deep learning segmentation algorithms in recent years. By comparing the segmentation effects of these three algorithms on various mineral thin-section images, the Gaussian pyramid segmentation algorithm based on bilateral filtering was selected as the optimal one. This was then combined with the directional maximum intercept method to measure the particle size of ilmenite and pyrite images. The experimental results show that the segmentation method based on the bilateral filtering Gaussian pyramid technique has higher segmentation accuracy than the other two algorithms, and can accurately measure the particle size of minerals under the microscope. Compared with manual measurement, this method can effectively and accurately measure the microscopic particle size of target minerals, greatly reducing the workload of measurement personnel and reducing the time spent on measurement. Full article
(This article belongs to the Section Mineral Processing and Extractive Metallurgy)
Show Figures

Figure 1

Figure 1
<p>Images of pyrite and lead–zinc ore under different shooting conditions; (<b>b</b>) lead–zinc ore, (<b>a</b>,<b>c</b>–<b>e</b>) are pyrite. The calibration scales of (<b>a</b>,<b>c</b>–<b>e</b>) are 0.2 mm, 100 um, 200 um, 500 um, and 1 mm, respectively.</p>
Full article ">Figure 2
<p>(<b>a</b>), (<b>b</b>), (<b>c</b>), (<b>d</b>), and (<b>e</b>) are the image segmentation results of the Gaussian pyramid based on bilateral filtering, corresponding to (<b>a1</b>), (<b>b1</b>), (<b>c1</b>), (<b>d1</b>), and (<b>e1</b>), respectively.</p>
Full article ">Figure 3
<p>The result of adding the mask image to (<b>a1</b>), (<b>b1</b>), (<b>c1</b>), (<b>d1</b>), and (<b>e1</b>) corresponds to (<b>a’1</b>), (<b>b’1</b>), (<b>c’1</b>), (<b>d’1</b>), and (<b>e’1</b>).</p>
Full article ">Figure 4
<p>(<b>a</b>), (<b>b</b>), (<b>c</b>), (<b>d</b>), and (<b>e</b>) use the mineral image segmentation results based on the FCM-WA algorithm to correspond to the results (<b>a2</b>), (<b>b2</b>), (<b>c2</b>), (<b>d2</b>) and (<b>e2</b>), respectively.</p>
Full article ">Figure 5
<p>Mask-R-CNN algorithm flow chart.</p>
Full article ">Figure 6
<p>(<b>a</b>), (<b>b</b>), (<b>c</b>), (<b>d</b>), and (<b>e</b>) are the segmentation results using Mask-R-CNN, corresponding to (<b>a3</b>), (<b>b3</b>), (<b>c3</b>), (<b>d3</b>), and (<b>e3</b>), respectively. Among them, Py is pyrite and Mag is lead–zinc ore.</p>
Full article ">Figure 7
<p>Comparison of edge segmentation using three algorithms in (<b>a</b>).</p>
Full article ">Figure 8
<p>Extraction of target particles from Gaussian pyramid segmentation results.</p>
Full article ">Figure 9
<p>The maximum intercept for the orientation of six particles.</p>
Full article ">Figure 10
<p>The extracted particles are marked and partially enlarged.</p>
Full article ">Figure 11
<p>Comparison of image measurement and manual measurement data; (<b>a</b>–<b>e</b>) are the data comparison between manual measurement and image measurement in <a href="#minerals-14-01284-f008" class="html-fig">Figure 8</a>a4–e4, respectively.</p>
Full article ">
39 pages, 4291 KiB  
Review
Machine Learning and Deep Learning for Crop Disease Diagnosis: Performance Analysis and Review
by Habiba Njeri Ngugi, Andronicus A. Akinyelu and Absalom E. Ezugwu
Agronomy 2024, 14(12), 3001; https://doi.org/10.3390/agronomy14123001 - 17 Dec 2024
Viewed by 305
Abstract
Crop diseases pose a significant threat to global food security, with both economic and environmental consequences. Early and accurate detection is essential for timely intervention and sustainable farming. This paper presents a review of machine learning (ML) and deep learning (DL) techniques for [...] Read more.
Crop diseases pose a significant threat to global food security, with both economic and environmental consequences. Early and accurate detection is essential for timely intervention and sustainable farming. This paper presents a review of machine learning (ML) and deep learning (DL) techniques for crop disease diagnosis, focusing on Support Vector Machines (SVMs), Random Forest (RF), k-Nearest Neighbors (KNNs), and deep models like VGG16, ResNet50, and DenseNet121. The review method includes an in-depth analysis of algorithm performance using key metrics such as accuracy, precision, recall, and F1 score across various datasets. We also highlight the data imbalances in commonly used datasets, particularly PlantVillage, and discuss the challenges posed by these imbalances. The research highlights critical insights regarding ML and DL models in crop disease detection. A primary challenge identified is the imbalance in the PlantVillage dataset, with a high number of healthy images and a strong bias toward certain disease categories like fungi, leaving other categories like mites and molds underrepresented. This imbalance complicates model generalization, indicating a need for preprocessing steps to enhance performance. This study also shows that combining Vision Transformers (ViTs) with Green Chromatic Coordinates and hybridizing these with SVM achieves high classification accuracy, emphasizing the value of advanced feature extraction techniques in improving model efficacy. In terms of comparative performance, DL architectures like ResNet50, VGG16, and convolutional neural network demonstrated robust accuracy (95–99%) across diverse datasets, underscoring their effectiveness in managing complex image data. Additionally, traditional ML models exhibited varied strengths; for instance, SVM performed better on balanced datasets, while RF excelled with imbalanced data. Preprocessing methods like K-means clustering, Fuzzy C-Means, and PCA, along with ensemble approaches, further improved model accuracy. Lastly, the study underscores that high-quality, well-labeled datasets, stakeholder involvement, and comprehensive evaluation metrics such as F1 score and precision are crucial for optimizing ML and DL models, making them more effective for real-world applications in sustainable agriculture. Full article
(This article belongs to the Collection Machine Learning in Digital Agriculture)
Show Figures

Figure 1

Figure 1
<p>Preferred reporting items for systematic reviews and meta-analysis (PRISMA) diagram for this study.</p>
Full article ">Figure 2
<p>General workflow of an ML-based crop detection technique.</p>
Full article ">Figure 3
<p>General workflow of a DL-based crop detection technique.</p>
Full article ">Figure 4
<p>Crop distribution in the PlantVillage dataset.</p>
Full article ">Figure 5
<p>Distribution of healthy and unhealthy samples in the PlantVillage dataset.</p>
Full article ">Figure 6
<p>Statistics of crop diseases in PlantVillage dataset.</p>
Full article ">Figure 7
<p>Classification accuracy of other ML algorithms from different authors.</p>
Full article ">Figure 8
<p>Classification accuracies of SVM-based crop detection techniques. Data sourced from [<a href="#B28-agronomy-14-03001" class="html-bibr">28</a>,<a href="#B29-agronomy-14-03001" class="html-bibr">29</a>,<a href="#B30-agronomy-14-03001" class="html-bibr">30</a>,<a href="#B31-agronomy-14-03001" class="html-bibr">31</a>,<a href="#B32-agronomy-14-03001" class="html-bibr">32</a>,<a href="#B33-agronomy-14-03001" class="html-bibr">33</a>].</p>
Full article ">Figure 9
<p>Classification accuracies of KNN-based crop detection techniques. Data sourced from [<a href="#B32-agronomy-14-03001" class="html-bibr">32</a>,<a href="#B34-agronomy-14-03001" class="html-bibr">34</a>,<a href="#B36-agronomy-14-03001" class="html-bibr">36</a>,<a href="#B37-agronomy-14-03001" class="html-bibr">37</a>].</p>
Full article ">Figure 10
<p>Classification accuracies of RF-based crop detection techniques. Data sourced from [<a href="#B32-agronomy-14-03001" class="html-bibr">32</a>,<a href="#B39-agronomy-14-03001" class="html-bibr">39</a>,<a href="#B40-agronomy-14-03001" class="html-bibr">40</a>,<a href="#B41-agronomy-14-03001" class="html-bibr">41</a>,<a href="#B42-agronomy-14-03001" class="html-bibr">42</a>].</p>
Full article ">Figure 11
<p>Classification accuracies of other ML-based crop detection techniques. Data sourced from [<a href="#B43-agronomy-14-03001" class="html-bibr">43</a>,<a href="#B44-agronomy-14-03001" class="html-bibr">44</a>,<a href="#B45-agronomy-14-03001" class="html-bibr">45</a>,<a href="#B46-agronomy-14-03001" class="html-bibr">46</a>].</p>
Full article ">Figure 12
<p>Performance of CNN-based crop detection techniques. Data sourced from [<a href="#B47-agronomy-14-03001" class="html-bibr">47</a>,<a href="#B48-agronomy-14-03001" class="html-bibr">48</a>,<a href="#B49-agronomy-14-03001" class="html-bibr">49</a>,<a href="#B50-agronomy-14-03001" class="html-bibr">50</a>,<a href="#B51-agronomy-14-03001" class="html-bibr">51</a>,<a href="#B52-agronomy-14-03001" class="html-bibr">52</a>,<a href="#B53-agronomy-14-03001" class="html-bibr">53</a>].</p>
Full article ">Figure 13
<p>Classification accuracies of other DL-based crop detection techniques. Data sourced from [<a href="#B46-agronomy-14-03001" class="html-bibr">46</a>,<a href="#B47-agronomy-14-03001" class="html-bibr">47</a>,<a href="#B49-agronomy-14-03001" class="html-bibr">49</a>,<a href="#B52-agronomy-14-03001" class="html-bibr">52</a>,<a href="#B53-agronomy-14-03001" class="html-bibr">53</a>,<a href="#B54-agronomy-14-03001" class="html-bibr">54</a>].</p>
Full article ">Figure 14
<p>Classification accuracies of VGG16-based crop detection techniques. Data sourced from [<a href="#B14-agronomy-14-03001" class="html-bibr">14</a>,<a href="#B56-agronomy-14-03001" class="html-bibr">56</a>,<a href="#B58-agronomy-14-03001" class="html-bibr">58</a>,<a href="#B60-agronomy-14-03001" class="html-bibr">60</a>,<a href="#B62-agronomy-14-03001" class="html-bibr">62</a>].</p>
Full article ">Figure 15
<p>Classification accuracies of ResNet-based crop detection techniques. Data sourced from [<a href="#B27-agronomy-14-03001" class="html-bibr">27</a>,<a href="#B53-agronomy-14-03001" class="html-bibr">53</a>,<a href="#B61-agronomy-14-03001" class="html-bibr">61</a>,<a href="#B63-agronomy-14-03001" class="html-bibr">63</a>,<a href="#B64-agronomy-14-03001" class="html-bibr">64</a>,<a href="#B65-agronomy-14-03001" class="html-bibr">65</a>].</p>
Full article ">Figure 16
<p>Classification accuracies of DenseNet121-based crop detection techniques. Data sourced from [<a href="#B10-agronomy-14-03001" class="html-bibr">10</a>,<a href="#B16-agronomy-14-03001" class="html-bibr">16</a>,<a href="#B18-agronomy-14-03001" class="html-bibr">18</a>,<a href="#B20-agronomy-14-03001" class="html-bibr">20</a>,<a href="#B46-agronomy-14-03001" class="html-bibr">46</a>,<a href="#B62-agronomy-14-03001" class="html-bibr">62</a>].</p>
Full article ">Figure 17
<p>Classification accuracies of the best-performing ML-based crop detection techniques [<a href="#B31-agronomy-14-03001" class="html-bibr">31</a>,<a href="#B35-agronomy-14-03001" class="html-bibr">35</a>,<a href="#B36-agronomy-14-03001" class="html-bibr">36</a>,<a href="#B43-agronomy-14-03001" class="html-bibr">43</a>].</p>
Full article ">Figure 18
<p>Comparison between the least-performing ML-based crop detection techniques.</p>
Full article ">Figure 19
<p>Classification accuracy of the best-performing DL-based techniques. Data sourced from [<a href="#B20-agronomy-14-03001" class="html-bibr">20</a>,<a href="#B49-agronomy-14-03001" class="html-bibr">49</a>,<a href="#B58-agronomy-14-03001" class="html-bibr">58</a>,<a href="#B62-agronomy-14-03001" class="html-bibr">62</a>].</p>
Full article ">
28 pages, 4684 KiB  
Article
Exploring the Role of Artificial Intelligence in Wastewater Treatment: A Dynamic Analysis of Emerging Research Trends
by Javier De la Hoz-M, Edwan Anderson Ariza-Echeverri and Diego Vergara
Resources 2024, 13(12), 171; https://doi.org/10.3390/resources13120171 - 16 Dec 2024
Viewed by 439
Abstract
Wastewater treatment is a critical process for ensuring water quality and public health, particularly in the context of increasing environmental challenges such as pollution and water scarcity. Artificial intelligence (AI) has emerged as a transformative technology capable of optimizing various wastewater treatment processes, [...] Read more.
Wastewater treatment is a critical process for ensuring water quality and public health, particularly in the context of increasing environmental challenges such as pollution and water scarcity. Artificial intelligence (AI) has emerged as a transformative technology capable of optimizing various wastewater treatment processes, such as contaminant removal, energy consumption, and cost-efficiency. This study presents a comprehensive bibliometric analysis of AI applications in wastewater treatment, utilizing data from Scopus and Web of Science covering 4335 publications from 1985 to 2024. Utilizing machine learning techniques such as neural networks, fuzzy logic, and genetic algorithms, the analysis reveals key trends in the role of the AI in optimizing wastewater treatment processes. The results show that AI has increasingly been applied to solve complex problems like membrane fouling, nutrient removal, and biofouling control. Regional contributions highlight a strong focus on advanced oxidation processes, microbial sludge treatment, and energy optimization. The Latent Dirichlet Allocation (LDA) model further identifies emerging topics such as real-time process monitoring and AI-driven effluent prediction as pivotal areas for future research. The findings provide valuable insights into the current state and future potential of AI technologies in wastewater management, offering a roadmap for researchers exploring the integration of AI to address sustainability challenges in the field. Full article
(This article belongs to the Special Issue Advances in Wastewater Reuse)
Show Figures

Figure 1

Figure 1
<p>PRISMA diagram illustrating the identification, screening, and selection of studies.</p>
Full article ">Figure 2
<p>Annual scientific publications and mean citations per article (1985–2024) related to AI in wastewater treatment.</p>
Full article ">Figure 3
<p>Geographical distribution of publications in AI-driven wastewater research (1985–2024).</p>
Full article ">Figure 4
<p>Top institutions contributing to AI-driven wastewater research (1985–2024).</p>
Full article ">Figure 5
<p>Collaboration network of countries in AI-driven wastewater research. This figure illustrates the global collaboration network, with node size representing the centrality and influence of each country. The connections depict collaborative ties between nations, with China serving as the dominant hub connecting various countries. This image can be best visualized in its HTML format, in <a href="#app1-resources-13-00171" class="html-app">Supplementary Material S1</a>.</p>
Full article ">Figure 6
<p>Collaboration network of institutions in AI-driven wastewater research. The figure illustrates the institutional collaboration network, with node size representing the influence and centrality of each institution. Colors correspond to different clusters, reflecting distinct communities within the global network. This image can be best visualized in its HTML format, in <a href="#app1-resources-13-00171" class="html-app">Supplementary Material S2</a>.</p>
Full article ">Figure 7
<p>Collaboration network of authors in AI-driven wastewater research. The figure illustrates the author collaboration network, with node size representing centrality and influence. Connections indicate collaborative ties, with prominent authors like Qiao J. and Wang Z. serving as major hubs in the global network. This image can be best visualized in its HTML format, in <a href="#app1-resources-13-00171" class="html-app">Supplementary Material S3</a>.</p>
Full article ">Figure 8
<p>Intertopic distance map of AI applications in wastewater management: LDA visualization using multidimensional scaling.</p>
Full article ">Figure 9
<p>Temporal evolution of research topics in AI-driven wastewater research (1985–2024).</p>
Full article ">Figure 10
<p>Heatmap of research topic distribution by country in AI-driven wastewater research.</p>
Full article ">Figure 11
<p>Heatmap of research topic distribution by journal in AI-driven wastewater research.</p>
Full article ">
20 pages, 6266 KiB  
Article
Temperature Control Strategy for Hydrogen Fuel Cell Based on IPSO-Fuzzy-PID
by Zenghui Liu, Haiying Dong and Xiping Ma
Electronics 2024, 13(24), 4949; https://doi.org/10.3390/electronics13244949 - 16 Dec 2024
Viewed by 306
Abstract
Hydrogen fuel cell water-thermal management systems suffer from slow response time, system vibration, and large temperature fluctuations of load current changes. In this paper, Logistic chaotic mapping, adaptively adjusted inertia weight and asymmetric learning factors are integrated to enhance the particle swarm optimization [...] Read more.
Hydrogen fuel cell water-thermal management systems suffer from slow response time, system vibration, and large temperature fluctuations of load current changes. In this paper, Logistic chaotic mapping, adaptively adjusted inertia weight and asymmetric learning factors are integrated to enhance the particle swarm optimization (PSO) algorithm and combine it with fuzzy control to propose an innovative improved particle swarm optimization-Fuzzy control strategy. The use of chaotic mapping to initialize the particle population effectively enhances the variety within the population, which subsequently improves the ability to search globally and prevents the algorithm from converging to a local optimum solution prematurely; by improving the parameters of learning coefficients and inertia weight, the global and local search abilities are balanced at different stages of the algorithm, so as to strengthen the algorithm’s convergence certainty while reducing the dependency on expert experience in fuzzy control. In this article, a fuel cell experimental platform is constructed to confirm the validity and efficiency of the recommended strategy, and the analysis reveals that the improved particle swarm optimization (IPSO) algorithm demonstrates better convergence performance than the standard PSO algorithm. The IPSO-Fuzzy-PID management approach is capable of providing a swift response and significantly diminishing the overshoot in the system’s performance, to maintain the system’s safe and stable execution. Full article
(This article belongs to the Section Systems & Control Engineering)
Show Figures

Figure 1

Figure 1
<p>Structure of fuel cell thermal management system.</p>
Full article ">Figure 2
<p>Flow chart of improved particle swarm algorithm.</p>
Full article ">Figure 3
<p>Structure of the Fuzzy-PID controller.</p>
Full article ">Figure 4
<p>Plot of error e affiliation function.</p>
Full article ">Figure 5
<p>(<b>a</b>) ∆Kp, (<b>b</b>) ∆Ki, (<b>c</b>) ∆<span class="html-italic">K</span><sub>d</sub> Surface Plot.</p>
Full article ">Figure 6
<p>Structure of IPSO-Fuzzy-PID controller.</p>
Full article ">Figure 7
<p>IPSO-Fuzzy-PID Simulink model diagram.</p>
Full article ">Figure 8
<p>Physical diagram of the fuel cell test system.</p>
Full article ">Figure 9
<p>Comparison of simulated and experimental values of polarization curves of the stack.</p>
Full article ">Figure 10
<p>Comparison of initialized particle distribution.</p>
Full article ">Figure 11
<p>Graph of learning factor dynamics.</p>
Full article ">Figure 12
<p>Convergence comparison curves, (<b>a</b>) F9 Objective function value curve, (<b>b</b>) F10 Objective function value curve.</p>
Full article ">Figure 13
<p>Iterative optimization curve, (<b>a</b>) <span class="html-italic">K</span><sub>p</sub> iterative optimization curve, (<b>b</b>) <span class="html-italic">K</span><sub>i</sub> iterative optimization curve, (<b>c</b>) <span class="html-italic">K</span><sub>d</sub> iterative optimization curve.</p>
Full article ">Figure 14
<p>Iterative optimization curve.</p>
Full article ">Figure 15
<p>Iterative optimization curve.</p>
Full article ">
30 pages, 8853 KiB  
Article
Research and Prediction Analysis of Key Factors Influencing the Carbon Dioxide Emissions of Countries Along the “Belt and Road” Based on Panel Regression and the A-A-E Coupling Model
by Xiang-Dong Feng, Xiang-Long Wang, Li Wen, Yao Yuan and Yu-Qin Zhang
Sustainability 2024, 16(24), 11014; https://doi.org/10.3390/su162411014 - 16 Dec 2024
Viewed by 374
Abstract
With the in-depth implementation of China’s “Belt and Road” strategic policy, member countries along the Belt and Road have gained enormous economic benefits. Thus, it is important to accurately grasp the factors that affect carbon emissions and coordinate the relationship between economic development [...] Read more.
With the in-depth implementation of China’s “Belt and Road” strategic policy, member countries along the Belt and Road have gained enormous economic benefits. Thus, it is important to accurately grasp the factors that affect carbon emissions and coordinate the relationship between economic development and environmental protection, which can impact the living environment of people worldwide. In this study, the researchers gathered data from the World Bank database, identified key indicators significantly impacting carbon emissions, employed the Pearson correlation coefficient and random forest model to perform dimensionality reduction on these indicators, and subsequently assessed the refined data using a panel regression model to examine the correlation and significance of these indicators and carbon emissions across various country types. To ensure the stability of the results, three prediction models were selected for coupling analysis: the adaptive neuro-fuzzy inference system (ANFIS) from the field of machine learning, the autoregressive integrated moving average (ARIMA) model, and the exponential smoothing method prediction model (ES) from the field of time series prediction. These models were used to assess carbon emissions from 54 countries along the Belt and Road from 2021 to 2030, and a coupling formula was defined to integrate the prediction results. The findings demonstrated that the integrated prediction amalgamates the forecasting traits of the three approaches, manifesting remarkable stability. The error analysis also indicated that the short-term prediction results are satisfactory. This has substantial practical implications for China in terms of fine-tuning its foreign policy, considering the entire situation and planning accordingly, and advancing energy conservation and emission reduction worldwide. Full article
Show Figures

Figure 1

Figure 1
<p>The distribution of countries along the Belt and Road.</p>
Full article ">Figure 2
<p>The changes in total carbon emissions based on panel data from the researched countries from 2005 to 2020.</p>
Full article ">Figure 3
<p>The changes in carbon emissions produced by four countries in specific years.</p>
Full article ">Figure 4
<p>Rates of change in carbon emissions of different countries based on panel data from 2005 to 2020. (<b>a</b>) Distribution of rates of change in carbon emissions in various countries in 2006. (<b>b</b>) Distribution of rates of change in carbon emissions in various countries in 2009. (<b>c</b>) Distribution of rates of change in carbon emissions in various countries in 2015. (<b>d</b>) Distribution of rates of change in carbon emissions in various countries in 2019.</p>
Full article ">Figure 5
<p>Annual rate of change in carbon emissions in Laos and Malta from 2005 to 2020: (<b>a</b>) Laos; (<b>b</b>) Malta.</p>
Full article ">Figure 6
<p>Data distribution and linear tests of indicator pairs.</p>
Full article ">Figure 7
<p>Data distribution before and after Box–Cox transformation for CNI and HTE.</p>
Full article ">Figure 8
<p>ANFIS network structure and calculation sequence.</p>
Full article ">Figure 9
<p>Results of the first and fifth error evaluation changes in the selection of random forest parameters.</p>
Full article ">Figure 10
<p>Correlation heatmap and importance ranking of indicators.</p>
Full article ">Figure 11
<p>Thermal map and significance criteria for coupling regression coefficients of mixed effect models for different types of countries.</p>
Full article ">Figure 12
<p>Comparison of true and predicted carbon emission data using multiple linear regression.</p>
Full article ">Figure 13
<p>PSO neural network used for predicting carbon emissions data: a comparison of actual and predicted values. (<b>a</b>) PSO network training set results. (<b>b</b>) PSO network test set results. (<b>c</b>) PSO algorithm fitness curve changes. (<b>d</b>) PSO neural network R<sup>2</sup> variation.</p>
Full article ">Figure 14
<p>Training effectiveness of the LSTM model for predicting carbon emissions. (<b>a</b>) Changes in MSE and loss during training. (<b>b</b>) LSTM carbon emission prediction situation.</p>
Full article ">Figure 15
<p>ANFIS prediction results of carbon emission data. (<b>a</b>) Comparison of ANFIS’s true and predicted carbon emissions. (<b>b</b>) Changes in the prediction error of ANFIS. (<b>c</b>) Histogram of ANFIS’s prediction error distribution.</p>
Full article ">Figure 16
<p>Cycle prediction flowchart.</p>
Full article ">Figure 17
<p>Comparison of parameter selection and prediction of the one-time exponential smoothing method model. (<b>a</b>) R<sup>2</sup> and RMSE corresponding to different model parameters. (<b>b</b>) Comparison of three parameter prediction results with actual values.</p>
Full article ">Figure 18
<p>Error and residual analysis of the triple exponential smoothing model’s parameter selection. (<b>a</b>) RMSE variation in different parameter combinations. (<b>b</b>) Residual variation in different parameter combinations.</p>
Full article ">Figure 19
<p>Partial results of the coupled model’s carbon emission prediction.</p>
Full article ">Figure 20
<p>Prediction effect diagram of 100 random errors.</p>
Full article ">
22 pages, 7903 KiB  
Article
Forecasting Forex Market Volatility Using Deep Learning Models and Complexity Measures
by Pavlos I. Zitis, Stelios M. Potirakis and Alex Alexandridis
J. Risk Financial Manag. 2024, 17(12), 557; https://doi.org/10.3390/jrfm17120557 - 13 Dec 2024
Viewed by 471
Abstract
In this article, we examine whether incorporating complexity measures as features in deep learning (DL) algorithms enhances their accuracy in predicting forex market volatility. Our approach involved the gradual integration of complexity measures alongside traditional features to determine whether their inclusion would provide [...] Read more.
In this article, we examine whether incorporating complexity measures as features in deep learning (DL) algorithms enhances their accuracy in predicting forex market volatility. Our approach involved the gradual integration of complexity measures alongside traditional features to determine whether their inclusion would provide additional information that improved the model’s predictive accuracy. For our analyses, we employed recurrent neural networks (RNNs), long short-term memory (LSTM), and gated recurrent units (GRUs) as DL model architectures, while using the Hurst exponent and fuzzy entropy as complexity measures. All analyses were conducted on intraday data from four highly liquid currency pairs, with volatility estimated using the Range-Based estimator. Our findings indicated that the inclusion of complexity measures as features significantly enhanced the accuracy of DL models in predicting volatility. In achieving this, we contribute to a relatively unexplored area of research, as this is the first instance of such an approach being applied to the prediction of forex market volatility. Additionally, we conducted a comparative analysis of the three models’ performance, revealing that the LSTM and GRU models consistently demonstrated a superior accuracy. Finally, our findings also have practical implications, as they may assist risk managers and policymakers in forecasting volatility in the forex market. Full article
(This article belongs to the Special Issue Machine Learning Applications in Finance, 2nd Edition)
Show Figures

Figure 1

Figure 1
<p>Schematic representation of simple RNN cell (<a href="#B66-jrfm-17-00557" class="html-bibr">Yu et al. 2019</a>).</p>
Full article ">Figure 2
<p>Schematic representation of LSTM cell (<a href="#B66-jrfm-17-00557" class="html-bibr">Yu et al. 2019</a>).</p>
Full article ">Figure 3
<p>Schematic representation of gated recurrent unit (GRU) cell (<a href="#B66-jrfm-17-00557" class="html-bibr">Yu et al. 2019</a>).</p>
Full article ">Figure 4
<p>Partitioning of the dataset and the grid search process: (<b>a</b>) the division of the dataset into training and test sets and (<b>b</b>) a schematic representation of grid search and cross-validation.</p>
Full article ">Figure 5
<p>Evolution of four forex market currency exchange rate prices (blue curves, left vertical axis) and corresponding Range-Based volatility (grey curves, right vertical axis) over the period from 28 August 2014, to 29 December 2023: (<b>a</b>) EUR/USD, (<b>b</b>) GBP/USD, (<b>c</b>) USD/CAD, and (<b>d</b>) USD/CHF. The vertical red lines delineate the data area utilized for model training (on the left) and the data area employed for model testing (on the right).</p>
Full article ">Figure 6
<p>Actual values of the Range-Based volatility for EUR/USD (grey curves) and the model predictions (colored curves) by feature case and DL model for the test dataset (i.e., over the period 8 April 2020, to 29 December 2023). More specifically, subfigures (<b>a</b>,<b>e</b>,<b>i</b>) show the actual values of the Range-Based volatility and the predictions of the RNN, LSTM, and GRU models, respectively, for Case I (where the feature used was Range-Based Volatility). Subfigures (<b>b</b>,<b>f</b>,<b>j</b>) show the actual values of the Range-Based volatility and the predictions of the RNN, LSTM, and GRU models, respectively, for Case II (where the features used were Range-Based Volatility, High, and Low). Subfigures (<b>c</b>,<b>g</b>,<b>k</b>) show the actual values of the Range-Based volatility and the predictions of the RNN, LSTM, and GRU models, respectively, for Case III (where the features used were Range-Based Volatility, Hurst Exponent, and <math display="inline"><semantics> <mrow> <mi>F</mi> <mi>u</mi> <mi>z</mi> <mi>z</mi> <mi>y</mi> <mi>E</mi> <mi>n</mi> </mrow> </semantics></math>). Subfigures (<b>d</b>,<b>h</b>,<b>l</b>) show the actual values of the Range-Based volatility and the predictions of the RNN, LSTM, and GRU models, respectively, for Case IV (where the features used were Range-Based Volatility, High, Low, Hurst Exponent, and <math display="inline"><semantics> <mrow> <mi>F</mi> <mi>u</mi> <mi>z</mi> <mi>z</mi> <mi>y</mi> <mi>E</mi> <mi>n</mi> </mrow> </semantics></math>).</p>
Full article ">Figure 7
<p>Actual values of the Range-Based volatility for GBP/USD (grey curves) and the model predictions (colored curves) by feature case and DL model for the test dataset (i.e., over the period 8 April 2020, to 29 December 2023). More specifically, subfigures (<b>a</b>,<b>e</b>,<b>i</b>) show the actual values of the Range-Based volatility and the predictions of the RNN, LSTM, and GRU models, respectively, for Case I (where the feature used was Range-Based Volatility). Subfigures (<b>b</b>,<b>f</b>,<b>j</b>) show the actual values of the Range-Based volatility and the predictions of the RNN, LSTM, and GRU models, respectively, for Case II (where the features used were Range-Based Volatility, High, and Low). Subfigures (<b>c</b>,<b>g</b>,<b>k</b>) show the actual values of the Range-Based volatility and the predictions of the RNN, LSTM, and GRU models, respectively, for Case III (where the features used were Range-Based Volatility, Hurst Exponent, and <math display="inline"><semantics> <mrow> <mi>F</mi> <mi>u</mi> <mi>z</mi> <mi>z</mi> <mi>y</mi> <mi>E</mi> <mi>n</mi> </mrow> </semantics></math>). Subfigures (<b>d</b>,<b>h</b>,<b>l</b>) show the actual values of the Range-Based volatility and the predictions of the RNN, LSTM, and GRU models, respectively, for Case IV (where the features used were Range-Based Volatility, High, Low, Hurst Exponent, and <math display="inline"><semantics> <mrow> <mi>F</mi> <mi>u</mi> <mi>z</mi> <mi>z</mi> <mi>y</mi> <mi>E</mi> <mi>n</mi> </mrow> </semantics></math>).</p>
Full article ">Figure 8
<p>Actual values of the Range-Based volatility for USD/CAD (grey curves) and the model predictions (colored curves) by feature case and DL model for the test dataset (i.e., over the period 8 April 2020, to 29 December 2023). More specifically, subfigures (<b>a</b>,<b>e</b>,<b>i</b>) show the actual values of the Range-Based volatility and the predictions of the RNN, LSTM, and GRU models, respectively, for Case I (where the feature used was Range-Based Volatility). Subfigures (<b>b</b>,<b>f</b>,<b>j</b>) show the actual values of the Range-Based volatility and the predictions of the RNN, LSTM, and GRU models, respectively, for Case II (where the features used were Range-Based Volatility, High, and Low). Subfigures (<b>c</b>,<b>g</b>,<b>k</b>) show the actual values of the Range-Based volatility and the predictions of the RNN, LSTM, and GRU models, respectively, for Case III (where the features used were Range-Based Volatility, Hurst Exponent, and <math display="inline"><semantics> <mrow> <mi>F</mi> <mi>u</mi> <mi>z</mi> <mi>z</mi> <mi>y</mi> <mi>E</mi> <mi>n</mi> </mrow> </semantics></math>). Subfigures (<b>d</b>,<b>h</b>,<b>l</b>) show the actual values of the Range-Based volatility and the predictions of the RNN, LSTM, and GRU models, respectively, for Case IV (where the features used were Range-Based Volatility, High, Low, Hurst Exponent, and <math display="inline"><semantics> <mrow> <mi>F</mi> <mi>u</mi> <mi>z</mi> <mi>z</mi> <mi>y</mi> <mi>E</mi> <mi>n</mi> </mrow> </semantics></math>).</p>
Full article ">Figure 9
<p>Actual values of the Range-Based volatility for USD/CHF (grey curves) and the model predictions (colored curves) by feature case and DL model for the test dataset (i.e., over the period 8 April 2020, to 29 December 2023). More specifically, subfigures (<b>a</b>,<b>e</b>,<b>i</b>) show the actual values of the Range-Based volatility and the predictions of the RNN, LSTM, and GRU models, respectively, for Case I (where the feature used was Range-Based Volatility). Subfigures (<b>b</b>,<b>f</b>,<b>j</b>) show the actual values of the Range-Based volatility and the predictions of the RNN, LSTM, and GRU models, respectively, for Case II (where the features used were Range-Based Volatility, High, and Low). Subfigures (<b>c</b>,<b>g</b>,<b>k</b>) show the actual values of the Range-Based volatility and the predictions of the RNN, LSTM, and GRU models, respectively, for Case III (where the features used were Range-Based Volatility, Hurst Exponent, and <math display="inline"><semantics> <mrow> <mi>F</mi> <mi>u</mi> <mi>z</mi> <mi>z</mi> <mi>y</mi> <mi>E</mi> <mi>n</mi> </mrow> </semantics></math>). Subfigures (<b>d</b>,<b>h</b>,<b>l</b>) show the actual values of the Range-Based volatility and the predictions of the RNN, LSTM, and GRU models, respectively, for Case IV (where the features used were Range-Based Volatility, High, Low, Hurst Exponent, and <math display="inline"><semantics> <mrow> <mi>F</mi> <mi>u</mi> <mi>z</mi> <mi>z</mi> <mi>y</mi> <mi>E</mi> <mi>n</mi> </mrow> </semantics></math>). A subsequent evaluation of the model performance on the test dataset was conducted using four statistical metrics (i.e., <math display="inline"><semantics> <mrow> <mi>M</mi> <mi>A</mi> <mi>E</mi> </mrow> </semantics></math>, <math display="inline"><semantics> <mrow> <mi>R</mi> <mi>M</mi> <mi>S</mi> <mi>E</mi> </mrow> </semantics></math>, <math display="inline"><semantics> <mrow> <mi>N</mi> <mi>R</mi> <mi>M</mi> <mi>S</mi> <mi>E</mi> </mrow> </semantics></math>, and <math display="inline"><semantics> <mrow> <mi>D</mi> <mi>S</mi> </mrow> </semantics></math>), as detailed in <a href="#sec3dot4-jrfm-17-00557" class="html-sec">Section 3.4</a>. This evaluation was performed for each of the three models, across all currency rates, and for each of the four feature sets. The results are presented in <a href="#jrfm-17-00557-t002" class="html-table">Table 2</a>.</p>
Full article ">
30 pages, 11752 KiB  
Article
Optimizing Outdoor Micro-Space Design for Prolonged Activity Duration: A Study Integrating Rough Set Theory and the PSO-SVR Algorithm
by Jingwen Tian, Zimo Chen, Lingling Yuan and Hongtao Zhou
Buildings 2024, 14(12), 3950; https://doi.org/10.3390/buildings14123950 - 12 Dec 2024
Viewed by 400
Abstract
This study proposes an optimization method based on Rough Set Theory (RST) and Particle Swarm Optimization–Support Vector Regression (PSO-SVR), aimed at enhancing the emotional dimension of outdoor micro-space (OMS) design, thereby improving users’ outdoor activity duration preferences and emotional experiences. OMS, as a [...] Read more.
This study proposes an optimization method based on Rough Set Theory (RST) and Particle Swarm Optimization–Support Vector Regression (PSO-SVR), aimed at enhancing the emotional dimension of outdoor micro-space (OMS) design, thereby improving users’ outdoor activity duration preferences and emotional experiences. OMS, as a key element in modern urban design, significantly enhances residents’ quality of life and promotes public health. Accurately understanding and predicting users’ emotional needs is the core challenge in optimizing OMS. In this study, the Kansei Engineering (KE) framework is applied, using fuzzy clustering to reduce the dimensionality of emotional descriptors, while RST is employed for attribute reduction to select five key design features that influence users’ emotions. Subsequently, the PSO-SVR model is applied to establish the nonlinear mapping relationship between these design features and users’ emotions, predicting the optimal configuration of OMS design. The results indicate that the optimized OMS design significantly enhances users’ intention to stay in the space, as reflected by higher ratings for emotional descriptors and increased preferences for longer outdoor activity duration, all exceeding the median score of the scale. Additionally, comparative analysis shows that the PSO-SVR model outperforms traditional methods (e.g., BPNN, RF, and SVR) in terms of accuracy and generalization for predictions. These findings demonstrate that the proposed method effectively improves the emotional performance of OMS design and offers a solid optimization framework along with practical guidance for future urban public space design. The innovative contribution of this study lies in the proposed data-driven optimization method that integrates machine learning and KE. This method not only offers a new theoretical perspective for OMS design but also establishes a scientific framework to accurately incorporate users’ emotional needs into the design process. The method contributes new knowledge to the field of urban design, promotes public health and well-being, and provides a solid foundation for future applications in different urban environments. Full article
(This article belongs to the Special Issue Art and Design for Healing and Wellness in the Built Environment)
Show Figures

Figure 1

Figure 1
<p>Fundamental concepts of RST.</p>
Full article ">Figure 2
<p>Schematic diagram of SVR.</p>
Full article ">Figure 3
<p>PSO-SVR flowchart.</p>
Full article ">Figure 4
<p>The proposed research framework.</p>
Full article ">Figure 5
<p>The 60 OMS samples on collection.</p>
Full article ">Figure 6
<p>Morphological deconstruction of OMS.</p>
Full article ">Figure 7
<p>The fitness curve of “sense of coziness”.</p>
Full article ">Figure 8
<p>The fitting diagram of “sense of coziness”.</p>
Full article ">Figure 9
<p>The prediction error on the test set.</p>
Full article ">Figure 10
<p>The fitting diagram of “sense of dynamism”, “sense of covertness”, and “sense of order”.</p>
Full article ">Figure 11
<p>The parameter results of the emotional descriptors.</p>
Full article ">Figure 12
<p>Design concept modeling of OMS.</p>
Full article ">Figure 13
<p>Comparison of scatter plot; each row represents the performance of four models on the same dataset.</p>
Full article ">Figure 14
<p>Evaluation of the design scheme.</p>
Full article ">
12 pages, 2009 KiB  
Article
Developing a Robust Fuzzy Inference Algorithm in a Dog Disease Pre-Diagnosis System for Casual Owners
by Kwang Baek Kim, Doo Heon Song and Hyun Jun Park
Animals 2024, 14(24), 3561; https://doi.org/10.3390/ani14243561 - 10 Dec 2024
Viewed by 296
Abstract
While the pet market is continuously rapidly increasing in Korea, pet dog owners feel uncomfortable in coping with pet dog’s health problems in time. In this paper, we propose a pre-diagnosis system based on neuro-fuzzy learning, enabling non-expert users to monitor their pets’ [...] Read more.
While the pet market is continuously rapidly increasing in Korea, pet dog owners feel uncomfortable in coping with pet dog’s health problems in time. In this paper, we propose a pre-diagnosis system based on neuro-fuzzy learning, enabling non-expert users to monitor their pets’ health by inputting observed symptoms. To develop such a system, we form a disease–symptom database based on several textbooks with veterinarians’ guidance and filtering. The system offers likely disease predictions and recommended coping strategies based on fuzzy inference. We evaluated three fuzzy inference algorithms—PFCM-R, FHAL, and MNFL. While PFCM-R achieved high accuracy with clean data, it struggled with noisy inputs. FHAL showed better noise tolerance but lower precision. PFCM-R is a variant of well-known fuzzy unsupervised learner FCM, and FHAL is the hybrid fuzzy inference engine based on Fuzzy Association Memory and a double-layered FCM we developed. To make the system more robust, we propose the multi-layered neuro-fuzzy learner (MNFL) in this paper, which effectively weakens the association strength between the disease and the observed symptoms, less related to the body part on which the abnormal symptoms are observed. In experiments that are designed to examine how the inference system reacts under increasing noisy input from the user, MNFL achieved 98% accuracy even with non-erroneous inputs, demonstrating superior robustness to other inference engines. This system empowers pet owners to detect health issues early, improving the quality of care and fostering more informed interactions with veterinarians, ultimately enhancing the well-being of companion animals. Full article
(This article belongs to the Section Human-Animal Interactions, Animal Behaviour and Emotion)
Show Figures

Figure 1

Figure 1
<p>The workflow of the inference system and ObservedPart selection interface.</p>
Full article ">Figure 2
<p>The fuzzy hybrid association learner structure.</p>
Full article ">Figure 3
<p>Multi-layered neuro-fuzzy learning.</p>
Full article ">Figure 4
<p>Connection strength <span class="html-italic">W</span><sub>1</sub>.</p>
Full article ">Figure 5
<p>Connection strength W<sub>2</sub>.</p>
Full article ">Figure 6
<p>Connection strength W<sub>3</sub> and relative distance of diseases.</p>
Full article ">Figure 7
<p>Accuracy trend over noise level.</p>
Full article ">Figure 8
<p>Top-three inference trend over noise level.</p>
Full article ">
21 pages, 4801 KiB  
Article
AI Dermatochroma Analytica (AIDA): Smart Technology for Robust Skin Color Classification and Segmentation
by Abderrachid Hamrani, Daniela Leizaola, Nikhil Kumar Reddy Vedere, Robert S. Kirsner, Kacie Kaile, Alexander Lee Trinidad and Anuradha Godavarty
Cosmetics 2024, 11(6), 218; https://doi.org/10.3390/cosmetics11060218 - 10 Dec 2024
Viewed by 1249
Abstract
Traditional methods for skin color classification, such as visual assessments and conventional image classification, face limitations in accuracy and consistency under varying conditions. To address this, we developed AI Dermatochroma Analytica (AIDA), an unsupervised learning system designed to enhance dermatological diagnostics. AIDA applies [...] Read more.
Traditional methods for skin color classification, such as visual assessments and conventional image classification, face limitations in accuracy and consistency under varying conditions. To address this, we developed AI Dermatochroma Analytica (AIDA), an unsupervised learning system designed to enhance dermatological diagnostics. AIDA applies clustering techniques to classify skin tones without relying on labeled data, evaluating over twelve models, including K-means, density-based, hierarchical, and fuzzy logic algorithms. The model’s key feature is its ability to mimic the process clinicians traditionally perform by visually matching the skin with the Fitzpatrick Skin Type (FST) palette scale but with enhanced precision and accuracy using Euclidean distance-based clustering techniques. AIDA demonstrated superior performance, achieving a 97% accuracy rate compared to 87% for a supervised convolutional neural network (CNN). The system also segments skin images into clusters based on color similarity, providing detailed spatial mapping aligned with dermatological standards. This segmentation reduces the uncertainty related to lighting conditions and other environmental factors, enhancing precision and consistency in skin color classification. This approach offers significant improvements in personalized dermatological care by reducing reliance on labeled data, improving diagnostic accuracy, and paving the way for future applications in diverse dermatological and cosmetic contexts. Full article
Show Figures

Figure 1

Figure 1
<p>Image transformation to the LAB color space.</p>
Full article ">Figure 2
<p>Systematic evaluation and visualization of cluster configurations.</p>
Full article ">Figure 3
<p>Illustrative scheme of the color matching methodology.</p>
Full article ">Figure 4
<p>Color alignment visualization between segmented skin colors and Fitzpatrick Skin Type palette.</p>
Full article ">Figure 5
<p>Sample of top-foot skin and FST scale palette imagery used in the comparative study of clustering models.</p>
Full article ">Figure 6
<p>Comparative visualization of clustering model performances in AIDA system.</p>
Full article ">Figure 7
<p>Bar chart for the evaluation of K-means AIDA vs. CNN performances in skin color classification.</p>
Full article ">Figure 8
<p>Confusion matrices for the comparative analysis of (<b>a</b>) AIDA and (<b>b</b>) CNN predicted outcomes vs. ground-truth FST classes.</p>
Full article ">Figure 9
<p>Skin regions using AIDA algorithm with two, three, and four cluster segments, matched to FST categories.</p>
Full article ">Figure A1
<p>Confusion matrix comparison of three ground-truth methods that could be used for testing the unsupervised convolutional network model from 48 images. Red coloration signifies a mismatch, while blue is a match for FST class. (<b>A</b>) Visual classification by a clinician versus visual classification by researcher, both using an FST sticker within image. (<b>B</b>) Commercial device FST classification using individual topological angle (ITA) measurements versus visual classification by researcher using an FST sticker within image. (<b>C</b>) Commercial device FST classification using individual topological angle (ITA) measurements versus visual classification by clinician using an FST sticker within image.</p>
Full article ">Figure A2
<p>Architecture of convolutional neural network (CNN) model used in the comparative analysis.</p>
Full article ">
18 pages, 529 KiB  
Article
eFC-Evolving Fuzzy Classifier with Incremental Clustering Algorithm Based on Samples Mean Value
by Emmanuel Tavares, Gray Farias Moita and Alisson Marques Silva
Big Data Cogn. Comput. 2024, 8(12), 183; https://doi.org/10.3390/bdcc8120183 - 6 Dec 2024
Viewed by 402
Abstract
This paper introduces a new multiclass classifier called the evolving Fuzzy Classifier (eFC). Starting its knowledge base from scratch, the eFC structure evolves based on a clustering algorithm that can add, merge, delete, or update clusters (= rules) simultaneously while providing class predictions. [...] Read more.
This paper introduces a new multiclass classifier called the evolving Fuzzy Classifier (eFC). Starting its knowledge base from scratch, the eFC structure evolves based on a clustering algorithm that can add, merge, delete, or update clusters (= rules) simultaneously while providing class predictions. The procedure to add clusters uses the procrastination idea to prevent outliers from affecting the quality of learning. Two pruning mechanisms are used to maintain a concise and compact structure. In the first, redundant clusters are merged based on a similarity measure, and in the second, obsolete and unrepresentative clusters are excluded based on an inactivity strategy. The center of the clusters is adjusted based on the mean value of the attributes. The eFC model was evaluated and compared with state-of-the-art evolving fuzzy systems on 8 randomly selected data streams from the UCI and Kaggle repositories. The experimental results indicate that the eFC outperforms or is at least comparable to alternative state-of-the-art models. Specifically, the eFC achieved an average accuracy of 7% to 37% higher than the competing classifiers. The results and comparisons demonstrate that the eFC is a promising alternative for classification tasks in non-stationary environments, offering good accuracy, a compact structure, low computational cost, and efficient processing time. Full article
Show Figures

Figure 1

Figure 1
<p>Clusters center update.</p>
Full article ">Figure 2
<p>Clusters merging.</p>
Full article ">Figure 3
<p>Scalability by incrementing the number of attributes (ms).</p>
Full article ">Figure 4
<p>Scalability by increasing the number of output classes (ms).</p>
Full article ">
23 pages, 1948 KiB  
Article
PerFuSIT: Personalized Fuzzy Logic Strategies for Intelligent Tutoring of Programming
by Konstantina Chrysafiadi and Maria Virvou
Electronics 2024, 13(23), 4827; https://doi.org/10.3390/electronics13234827 - 6 Dec 2024
Viewed by 409
Abstract
Recent advancements in intelligent tutoring systems (ITS) driven by artificial intelligence (AI) have attracted substantial research interest, particularly in the domain of computer programming education. Given the diversity in learners’ backgrounds, cognitive abilities, and learning paces, the development of personalized tutoring strategies to [...] Read more.
Recent advancements in intelligent tutoring systems (ITS) driven by artificial intelligence (AI) have attracted substantial research interest, particularly in the domain of computer programming education. Given the diversity in learners’ backgrounds, cognitive abilities, and learning paces, the development of personalized tutoring strategies to support the effective attainment of learning objectives has become a critical challenge. This paper introduces personalized fuzzy logic strategies for intelligent programming tutoring (PerFuSIT), an innovative fuzzy logic-based module designed to select the most appropriate tutoring strategy from five available options, based on individual learner characteristics. The available strategies include revisiting previous content, progressing to the next topic, providing supplementary materials, assigning additional exercises, or advising the learner to take a break. PerFuSIT’s decision-making process incorporates a range of learner-specific parameters, such as performance metrics, error types, indicators of carelessness, frequency of help requests, and the time required to complete tasks. Embedded within the traditional ITS framework, PerFuSIT introduces a sophisticated reasoning mechanism for dynamically determining the optimal instructional approach. Experimental evaluations demonstrate that PerFuSIT significantly enhances learner performance and improves the overall efficacy of interactions with the ITS. The findings highlight the potential of fuzzy logic to optimize adaptive tutoring strategies by customizing instruction to individual learners’ strengths and weaknesses, thereby providing more effective and personalized educational support in programming instruction. Full article
Show Figures

Figure 1

Figure 1
<p>ITS that incorporates PerFuSIT.</p>
Full article ">Figure 2
<p>Architecture of PerFuSIT.</p>
Full article ">Figure A1
<p>Fuzzy sets for tutoring strategy significance.</p>
Full article ">Figure A2
<p>Fuzzy sets for input variables of PerFuSIT.</p>
Full article ">
25 pages, 6353 KiB  
Article
Fractional-Order Controller for the Course Tracking of Underactuated Surface Vessels Based on Dynamic Neural Fuzzy Model
by Guangyu Li, Yanxin Li, Xiang Li, Mutong Liu, Xuesong Zhang and Hua Jin
Fractal Fract. 2024, 8(12), 720; https://doi.org/10.3390/fractalfract8120720 - 5 Dec 2024
Viewed by 461
Abstract
Aiming at the uncertainty problem caused by the time-varying modeling parameters associated with ship speed in the course tracking control of underactuated surface vessels (USVs), this paper proposes a control algorithm based on the dynamic neural fuzzy model (DNFM). The DNFM simultaneously adjusts [...] Read more.
Aiming at the uncertainty problem caused by the time-varying modeling parameters associated with ship speed in the course tracking control of underactuated surface vessels (USVs), this paper proposes a control algorithm based on the dynamic neural fuzzy model (DNFM). The DNFM simultaneously adjusts the structure and parameters during learning and fully approximates the inverse dynamics of ships. Online identification and modeling lays the model foundation for ship motion control. The trained DNFM, serving as an inverse controller, is connected in parallel with the fractional-order PIλDμ controller to be used for the tracking control of the ship’s course. Moreover, the weights of the model can be further adjusted during the course tracking. Taking the actual ship data of a 5446 TEU large container ship, simulation experiments are conducted, respectively, for course tracking, course tracking under wind and wave interferences, and comparison with five different controllers. This proposed controller can overcome the influence of the uncertainty of modeling parameters, tracking the desired course quickly and effectively. Full article
(This article belongs to the Special Issue Applications of Fractional-Order Systems to Automatic Control)
Show Figures

Figure 1

Figure 1
<p>Motion coordinate system.</p>
Full article ">Figure 2
<p>Corresponding nonlinear ship model.</p>
Full article ">Figure 3
<p>Dynamic neural fuzzy model structure.</p>
Full article ">Figure 4
<p>Identification process of inverse model for ship course control.</p>
Full article ">Figure 5
<p>Flow of inverse model identification for ship course control based on DNFM.</p>
Full article ">Figure 6
<p>Ship course control system.</p>
Full article ">Figure 7
<p>Change in ship speed V.</p>
Full article ">Figure 8
<p>Change in ship model parameters <math display="inline"><semantics> <mrow> <mi mathvariant="normal">K</mi> </mrow> </semantics></math>, <math display="inline"><semantics> <mrow> <mi mathvariant="normal">T</mi> </mrow> </semantics></math>, <math display="inline"><semantics> <mrow> <mi mathvariant="sans-serif">α</mi> </mrow> </semantics></math>, and <math display="inline"><semantics> <mrow> <mi mathvariant="sans-serif">β</mi> </mrow> </semantics></math>.</p>
Full article ">Figure 9
<p>DNFM generating fuzzy rules.</p>
Full article ">Figure 10
<p>DNFM identification results.</p>
Full article ">Figure 11
<p>Root mean squared error in learning.</p>
Full article ">Figure 12
<p>DNFM identification error.</p>
Full article ">Figure 13
<p>Ship course tracking.</p>
Full article ">Figure 14
<p>Rudder control for ship course.</p>
Full article ">Figure 15
<p>Equivalent rudder angle of wind.</p>
Full article ">Figure 16
<p>Equivalent rudder angle of waves.</p>
Full article ">Figure 17
<p>DNFM generating fuzzy rules under wind and wave disturbances.</p>
Full article ">Figure 18
<p>DNFM identification results under wind and wave disturbances.</p>
Full article ">Figure 19
<p>Root mean squared error under wind and wave disturbances.</p>
Full article ">Figure 20
<p>Identification error of DNFM.</p>
Full article ">Figure 21
<p>Course control and rudder angle curves under wind and wave disturbances.</p>
Full article ">Figure 22
<p>Comparison of five different controllers.</p>
Full article ">Figure 23
<p>Rudder angle using five different controllers.</p>
Full article ">
21 pages, 59527 KiB  
Article
Recognition and Classification of Typical Building Shapes Based on YOLO Object Detection Models
by Xiao Wang, Haizhong Qian, Limin Xie, Xu Wang and Bohao Li
ISPRS Int. J. Geo-Inf. 2024, 13(12), 433; https://doi.org/10.3390/ijgi13120433 - 2 Dec 2024
Viewed by 538
Abstract
The recognition and classification of building shapes are the prerequisites and foundation for building simplification, matching, and change detection, which have always been important research problems in the field of cartographic generalization. Due to the ambiguity and uncertainty of building shape outlines, it [...] Read more.
The recognition and classification of building shapes are the prerequisites and foundation for building simplification, matching, and change detection, which have always been important research problems in the field of cartographic generalization. Due to the ambiguity and uncertainty of building shape outlines, it is difficult to describe them using unified rules, which has always limited the quality and automation level of building shape recognition. In response to the above issues, by introducing object detection technology in computer vision, this article proposes a building shape recognition and classification method based on the YOLO object detection model. Firstly, for different types of buildings, four levels of building training data samples are constructed, and YOLOv5, YOLOv8, YOLOv9, and YOLOv9 integrating attention modules are selected for training. The trained models are used to test the shape judgment of buildings in the dataset and verify the learning effectiveness of the models. The experimental results show that the YOLO model can accurately classify and locate the shape of buildings, and its recognition and detection effect have the ability to simulate advanced human visual cognition, which provides a new solution for the fuzzy shape recognition of buildings with complex outlines and local deformation. Full article
Show Figures

Figure 1

Figure 1
<p>Typical types of building shapes.</p>
Full article ">Figure 2
<p>Network architecture of YOLOv9.</p>
Full article ">Figure 3
<p>The GELAN module of YOLOv9 [<a href="#B46-ijgi-13-00433" class="html-bibr">46</a>].</p>
Full article ">Figure 4
<p>The introduction position of attention modules.</p>
Full article ">Figure 5
<p>The introduction position of attention modules [<a href="#B49-ijgi-13-00433" class="html-bibr">49</a>].</p>
Full article ">Figure 6
<p>The introduction position of attention modules [<a href="#B50-ijgi-13-00433" class="html-bibr">50</a>].</p>
Full article ">Figure 7
<p>The structures of CAM and SAM in the NAM module [<a href="#B51-ijgi-13-00433" class="html-bibr">51</a>].</p>
Full article ">Figure 8
<p>The structure of the SimAM attention module [<a href="#B52-ijgi-13-00433" class="html-bibr">52</a>].</p>
Full article ">Figure 9
<p>Recognition procedure of building shape based on YOLO object detection model.</p>
Full article ">Figure 10
<p>Examples of test datasets: (<b>a</b>) single building; (<b>b</b>) complex scene; (<b>c</b>) large-area scene.</p>
Full article ">Figure 11
<p>Misidentification examples of building shape: (<b>a</b>) F-like; (<b>b</b>) cross-like.</p>
Full article ">Figure 12
<p>Detection results of YOLOv9e on Test Dataset 2: (<b>a</b>) standard shape; (<b>b</b>) complex contour; (<b>c</b>) local deformation; (<b>d</b>) complex contour combined with local deformation.</p>
Full article ">Figure 13
<p>Simulation of advanced human visual cognition by YOLO models: (<b>a</b>) abstract summarizing; (<b>b</b>) edge detection; (<b>c</b>) fuzzy judgment; (<b>d</b>) local recognition; (<b>e</b>) analogical reasoning; (<b>f</b>) visual extension.</p>
Full article ">Figure 14
<p>Comparison between (<b>a</b>) GNN method and (<b>b</b>) YOLOv9e + CBAM with complex shapes.</p>
Full article ">
14 pages, 674 KiB  
Protocol
A Teamwork-Based Protocol for a Holistic Approach to Selecting a Sustainable Mine Dewatering Management Plan
by Dragoljub Bajić, Sanja Bajić, Jelena Trivan, Ljubica Figun and Jelena Milovanović
Sustainability 2024, 16(23), 10424; https://doi.org/10.3390/su162310424 - 28 Nov 2024
Viewed by 430
Abstract
The primary objective of the protocol is to establish and develop several scientific methodological procedures applicable to the design and selection of a suitable mine dewatering management plan. A significant challenge and contribution of the research lies in the initial hypothesis, which posits [...] Read more.
The primary objective of the protocol is to establish and develop several scientific methodological procedures applicable to the design and selection of a suitable mine dewatering management plan. A significant challenge and contribution of the research lies in the initial hypothesis, which posits the feasibility of organizing a multidisciplinary team to collaboratively determine the optimal solution for long-term mine dewatering. Protection against groundwater is a highly complex hydrogeological challenge, particularly in mining operations. Mines are inherently dynamic systems, constantly expanding both horizontally and vertically, from the very beginning of mining, also reaching significant depths. Given the inherent uncertainty in geologic systems, such as ore deposits, the entire dewatering process requires continuous “learning” and hierarchical problem-solving. Addressing these complexities involved forming a team of experts, leveraging their knowledge and experience, as well as several methodological procedures based on applied mathematics in geosciences and mining engineering, such as numerical modeling and simulation, fuzzy optimization and decision analysis. These circumstances necessitated continual adjustment to evolving operating conditions and prompted the development of a protocol for effective dewatering planning and mineral ore protection against groundwater. Such a protocol generates alternative mine dewatering solutions and considers their individual characteristics. Additionally, it defines and analyzes multiple criteria for evaluating the solutions and selecting a method that ensures optimal decision-making. The applied methods constitute a holistic system, represented by a single protocol, which includes an interdisciplinary approach to creating sustainable groundwater management strategies. Full article
(This article belongs to the Special Issue Sustainable Mining and Circular Economy)
Show Figures

Figure 1

Figure 1
<p>Protocol for designing and selecting the optimal mine dewatering management plan.</p>
Full article ">
Back to TopTop