Journal of Intelligent & Fuzzy Systems - Volume 45, issue 5

Authors: Cho, Seung-Beom | Jeong, Si-Hwa | Yu, Jae-Wook | Choi, Jae-Boong | Kim, Moon Ki

Article Type: Research Article

Abstract: Despite the significant improvements in the detection and diagnosis of plant diseases at an early stage facilitated by deep learning technology, there are challenges associated with the generalization performance of deep learning models. These problems from the differences between in-field and in-lab data, as well as the heterogeneity of training and prediction data features. In the case of tomato leaf diseases, the PlantVillage dataset is widely used and has already demonstrated accuracy of more than 99%. However, using trained model based on this dataset to predict in-field data results in low accuracy due to domain differences and heterogeneous features. In …this paper, we propose a domain adaptation method based on CycleGAN to solve this problem, followed by a preprocessing technique that utilizes both the OpenCV module and a segmentation model based on U-Net for the best generalization performance. The classification accuracy is evaluated by applying the DenseNet121 model trained on the PlantVillage dataset to the images generated by CycleGAN. Our results demonstrate, with an F1-score of 95.6%, that our domain adaptation method between the two domains is effective in mitigating the effect of domain shift. Show more

Keywords: Image processing, leaf classification, deep learning, CycleGAN, domain adaptation, tomato leaf disease

DOI: 10.3233/JIFS-230561

Citation: Journal of Intelligent & Fuzzy Systems, vol. 45, no. 5, pp. 8859-8870, 2023

Authors: Wang, Ru | Peng, Kexin | Liu, Fang | Li, Shugang

Article Type: Research Article

Abstract: With the increasing of online social behavior, social relationships have an important impact on consumer negative comment behavior (CNCB) on social commerce platforms. Existing studies lack to describe CNCB influenced by social relationships on social commerce platforms from the perspective of well-thought-out planning results, and the proposed structural equation models in previous studies have been difficult to predict CNCB. Hence, this study proposes a new structural equation model (SEM) and artificial neural network (ANN) model to deeply explore and reveal the generation mechanism of CNCB in the context of social commerce platforms based on the theory of planned behavior (TPB). …We regard social support as a moderating effect and construct a consumer negative comment planning behavior model (CNCPBM). The results of the data analysis show CNCPBM is supported. This study provides an important theoretical and practical contribution to CNCB, and offers practical management enlightenment for the managers of social commerce platforms. Show more

Keywords: Social commerce platforms, theory of planned behavior, artificial neural network, social support, negative comment behavior

DOI: 10.3233/JIFS-230563

Citation: Journal of Intelligent & Fuzzy Systems, vol. 45, no. 5, pp. 8871-8888, 2023

Authors: Wu, Meiqin | Yang, Jindou | Fan, Jianping

Article Type: Research Article

Abstract: With the continuous improvement and development of various decision-making methods, it has led to the widespread use of fuzzy sets and fuzzy numbers. At the same time, the application of decision-making methods in different fuzzy environments has been very effective in addressing the deficiencies in existing research. At present, triangular fuzzy numbers have been widely used in the evaluation aspects of various decision making methods, and the proposed R-number effectively solve the uncertainty involving problems related to future events, but the existing research based on the TOPSIS method in the R-number environment has not yet been clearly applied to the …triangular fuzzy number environment, and the indifference threshold-based attribute ratio analysis (ITARA) method in the fuzzy environment has yet to be extended. Therefore, this paper proposes a fuzzy indifference threshold-based attribute ratio analysis (FITARA) method based on triangular fuzzy numbers for solving the problem of determining attribute weights in the multi-attribute decision-making process. Secondly, the various risks of the decision environment and the impact on future events are considered and R-number are used to solve this puzzle. In addition, the incorporation of risk perception factors in the context of the existing RTOPSIS method considering multiple risk factors and the use of Manhattan distances to optimize the large number of operations in the process of the method resulted in the development of the FITARA-RTOPSIS model. Finally, the proposed FITARA-RTOPSIS method is applied to the problem of siting emergency supplies storage depots, and the effectiveness of the proposed method is verified through comparative analysis. Show more

Keywords: FITARA, R-number, RTOPSIS, Manhattan distance, TFN

DOI: 10.3233/JIFS-232393

Citation: Journal of Intelligent & Fuzzy Systems, vol. 45, no. 5, pp. 8889-8905, 2023

Authors: Shyama, S. | Iyer, Radha R.

Article Type: Research Article

Abstract: The attractive properties of the hypercube graph such as its diameter, good connectivity, and symmetry have made it a popular topology for the design of multi-computer interconnection networks. Efforts to improve some of these properties have led to the evolution of hypercube variants. Let c be the proper coloring of graph G , where the neighboring vertices will get individual colors. Coloring c is irregular if distinct vertices have distinct color codes and the least number of colors that ought to receive an irregular coloring is the irregular chromatic number, χir  (G ). In this paper, we …discuss the irregular coloring and find the irregular chromatic number for the hypercube graph Q n and some of its variants using binomial coefficients for the Locally twisted cube graph LTQ n , Crossed cube graph CQ n and two types of Fractal cubic network graph FCNG 1  (k ) and FCNG 2  (k ). Show more

Keywords: Irregular coloring, irregular chromatic number, hypercube graph, variants of hypercube graph

DOI: 10.3233/JIFS-232471

Citation: Journal of Intelligent & Fuzzy Systems, vol. 45, no. 5, pp. 8907-8913, 2023

Authors: Prabhakaran, Sudarsan | Ayyamperumal, Niranjil Kumar

Article Type: Research Article

Abstract: This manuscript proposes an automated artifacts detection and multimodal classification system for human emotion analysis from human physiological signals. First, multimodal physiological data, including the Electrodermal Activity (EDA), electrocardiogram (ECG), Blood Volume Pulse (BVP) and respiration rate signals are collected. Second, a Modified Compressed Sensing-based Decomposition (MCSD) is used to extract the informative Skin Conductance Response (SCR) events of the EDA signal. Third, raw features (edge and sharp variations), statistical and wavelet coefficient features of EDA, ECG, BVP, respiration and SCR signals are obtained. Fourth, the extracted raw features, statistical and wavelet coefficient features from all physiological signals are fed …into the parallel Deep Convolutional Neural Network (DCNN) to reduce the dimensionality of feature space by removing artifacts. Fifth, the fused artifact-free feature vector is obtained for neutral, stress and pleasure emotion classes. Sixth, an artifact-free feature vector is used to train the Random Forest Deep Neural Network (RFDNN) classifier. Then, a trained RFDNN classifier is applied to classify the test signals into different emotion classes. Thus, leveraging the strengths of both RF and DNN algorithms, more comprehensive feature learning using multimodal psychological data is achieved, resulting in robust and accurate classification of human emotional activities. Finally, an extensive experiment using the Wearable Stress and Affect Detection (WESAD) dataset shows that the proposed system outperforms other existing human emotion classification systems using physiological data. Show more

Keywords: Emotional reactivity, physiological signals, modified compressed sensing, motion artifacts, deep convolutional neural network, random forest deep neural network

DOI: 10.3233/JIFS-232662

Citation: Journal of Intelligent & Fuzzy Systems, vol. 45, no. 5, pp. 8915-8929, 2023

Authors: Zhao, Yiming | Zhao, Hongdong | Zhang, Xuezhi | Liu, Weina

Article Type: Research Article

Abstract: In Intelligent Transport Systems (ITS), vision is the primary mode of perception. However, vehicle images captured by low-cost traffic cameras under challenging weather conditions often suffer from poor resolution and insufficient detail representation. On the other hand, vehicle noise provides complementary auditory features that offer advantages such as environmental adaptability and a large recognition distance. To address these limitations and enhance the accuracy of low-quality traffic surveillance classification and identification, an effective audio-visual feature fusion method is crucial. This paper presents a research study that establishes an Urban Road Vehicle Audio-visual (URVAV) dataset specifically designed for low-quality images and noise …recorded in complex weather conditions. For low-quality vehicle image classification, the paper proposes a simple Convolutional Neural Network (CNN)-based model called Low-quality Vehicle Images Net (LVINet). Additionally, to further enhance classification accuracy, a spatial channel attention-based audio-visual feature fusion method is introduced. This method converts one-dimensional acoustic features into a two-dimensional audio Mel-spectrogram, allowing for the fusion of auditory and visual features. By leveraging the high correlation between these features, the representation of vehicle characteristics is effectively enhanced. Experimental results demonstrate that LVINet achieves a classification accuracy of 93.62% with reduced parameter count compared to existing CNN models. Furthermore, the proposed audio-visual feature fusion method improves classification accuracy by 7.02% and 4.33% when compared to using single audio or visual features alone, respectively. Show more

Keywords: Vehicle classification, feature fusion, convolutional neural network, low-quality images

DOI: 10.3233/JIFS-232812

Citation: Journal of Intelligent & Fuzzy Systems, vol. 45, no. 5, pp. 8931-8944, 2023

Authors: Chen, Hao | Su, Ze | Xu, Xiangqian

Article Type: Research Article

Abstract: The rapid development of global information technology, especially the emergence and widespread application of the Internet, has enabled information technology to quickly penetrate into various fields of the economy and society. Informatization and networking have become important features of today’s era. However, while people enjoy the tremendous progress brought by information technology to humanity, the openness and security vulnerabilities of computer networks have also made network information security issues increasingly prominent. The invasion of hackers, the continuous generation and spread of computer virus, and the rampant use of rogue software have all caused great economic losses to individuals, enterprises, and …countries. The computer network security evaluation is a multiple-attribute group decision making (MAGDM). Then, the TODIM and TOPSIS method has been established to deal with MAGDM issues. The interval neutrosophic sets (INSs) are established as an effective tool for representing uncertain information during the computer network security evaluation. In this manuscript, the interval neutrosophic number TODIM-TOPSIS (INN-TODIM-TOPSIS) method is established to solve the MAGDM under INSs. Finally, a numerical example study for computer network security evaluation is established to validate the INN-TODIM-TOPSIS method. The main research contribution of this paper is established: (1) the INN-TODIM-TOPSIS method is put up for MAGDM with INSs; (2) the INN-TODIM-TOPSIS method is put up for computer network security evaluation and were compared with existing methods; (3) Through the detailed comparison, it is evident that INN-TODIM-TOPSIS method for computer network security evaluation proposed in this paper are effective. Show more

Keywords: Multiple-attribute group decision making (MAGDM), Interval neutrosophic sets (INSs), TODIM method, TOPSIS method, Computer network security evaluation

DOI: 10.3233/JIFS-233181

Citation: Journal of Intelligent & Fuzzy Systems, vol. 45, no. 5, pp. 8945-8957, 2023

Authors: Wang, Yixuan | Zhang, Xiaowen

Article Type: Research Article

Abstract: The lean management and innovation capability evaluation of technological small and medium sized enterprises is a classical multi-attributes group decision-making (MAGDM). Recently, the probabilistic hesitant fuzzy sets (PHFSs) have been extended to apply in many fields. However, the existing models don’t evaluate the alternative considering the psychological factors. Thus, in this paper, an extended probabilistic hesitant fuzzy grey relational analysis (PHF-GRA) method is proposed to reduce the restrictions of GRA method by combining with cumulative prospect theory (CPT), considering the psychological preference. In addition, the PHFSs assigns probability values to different degrees of hesitancy, which shows its superiority in complex …environment. At the same time, the weight vectors of each attribute are calculated by the entropy values of different foreground decision elements. Then, probabilistic hesitant fuzzy GRA (PHF-GRA) model based on CPT model is constructed for MAGDM under PHFSs. Finally, a practical example study for lean management and innovation capability evaluation of technological small and medium sized enterprises is constructed to validate the proposed GRA (PHF-GRA) model based on model CPT and some comparative studies are constructed to verify the applicability. Show more

Keywords: Multi-attributes group decision-making (MAGDM), probabilistic hesitant fuzzy sets (PHFSs), grey relational analysis (GRA) method, entropy, lean management and innovation capability evaluation

DOI: 10.3233/JIFS-233403

Citation: Journal of Intelligent & Fuzzy Systems, vol. 45, no. 5, pp. 8959-8972, 2023

Authors: Wang, Yahui | Chen, Hongchang | Liu, Shuxin | Li, Xing | Hu, Yuxiang

Article Type: Research Article

Abstract: With the continuous escalation of telecommunication fraud modes, telecommunication fraud is becoming more and more concealed and disguised. Existing Graph Neural Networks (GNNs)-based fraud detection methods directly aggregate the neighbor features of target nodes as their own updated features, which preserves the commonality of neighbor features but ignores the differences with target nodes. This makes it difficult to effectively distinguish fraudulent users from normal users. To address this issue, a new model named Feature Difference-aware Graph Neural Network (FDAGNN) is proposed for detecting telecommunication fraud. FDAGNN first calculates the feature differences between target nodes and their neighbors, then adopts GAT …method to aggregate these feature differences, and finally uses GRU approach to fuse the original features of target nodes and the aggregated feature differences as the updated features of target nodes. Extensive experiments on two real-world telecom datasets demonstrate that FDAGNN outperforms seven baseline methods in the majority of metrics, with a maximum improvement of about 5%. Show more

Keywords: Fraud detection, graph neural networks, telecommunication networks, feature fusion

DOI: 10.3233/JIFS-221893

Citation: Journal of Intelligent & Fuzzy Systems, vol. 45, no. 5, pp. 8973-8988, 2023

Authors: Yuvaraj, S. | Vijay Franklin, J.

Article Type: Research Article

Abstract: The predictions of cognitive emotions are complex due to various cognitive emotion modalities. Deep network model has recently been used with huge cognitive emotion determination. The visual and auditory modalities of cognitive emotion recognition system are proposed. The extraction of powerful features helps obtain the content related to cognitive emotions for different speaking styles. Convolutional neural network (CNN) is utilized for feature extraction from the speech. On the other hand, the visual modality uses the 50 layers of a deep residual network for prediction purpose. Also, extracting features is important as the datasets are sensitive to outliers when trying to …model the content. Here, a long short-term memory network (LSTM) is considered to manage the issue. Then, the proposed Dense Layer Model (DLM) is trained in an E2E manner based on feature correlation that provides better performance than the conventional techniques. The proposed model gives 99% prediction accuracy which is higher to other approaches. Show more

Keywords: Cognitive emotion recognition, deep learning, prediction, visual modality, handcrafted features

DOI: 10.3233/JIFS-230766

Citation: Journal of Intelligent & Fuzzy Systems, vol. 45, no. 5, pp. 8989-9005, 2023