Intelligent Systems with Applications最新文献

Knowledge graphs (KGs) serve as invaluable tools for organizing and representing structural information, enabling powerful data analysis and retrieval. In this paper, we propose a novel knowledge graph learning algorithm based on deep convolutional neural networks (KGLA-DCNN) to enhance the classification accuracy of KG nodes. Leveraging the hierarchical and relational nature of KGs, our algorithm utilizes deep convolutional neural networks to capture intricate patterns and dependencies within the graph. We evaluate the effectiveness of KGLA-DCNN on two benchmark datasets, Cora and Citeseer, renowned for their challenging node classification tasks. Through extensive experiments, we demonstrate that our proposed algorithm significantly improves classification accuracy compared to state-of-the-art methods, showcasing its capability to leverage the rich structural information inherent in KGs. The results highlight the potential of deep convolutional neural networks in enhancing the learning and representation capabilities of knowledge graphs, paving the way for more accurate and efficient knowledge discovery in diverse domains.

Sign language (SL) serves as a visual communication tool bearing great significance for deaf people to interact with others and facilitate their daily life. Wide varieties of SLs and the lack of interpretation knowledge necessitate developing automated sign language recognition (SLR) systems to attenuate the communication gap between the deaf and hearing communities. Despite numerous advanced static SLR systems, they are not practical and favorable enough for real-life scenarios once assessed simultaneously from different critical aspects: accuracy in dealing with high intra- and slight inter-class variations, robustness, computational complexity, and generalization ability. To this end, we propose a novel multi-lingual multi-modal SLR system, namely MLMSign, by taking full strengths of hand-crafted features and deep learning models to enhance the performance and the robustness of the system against illumination changes while minimizing computational cost. The RGB sign images and 2D visualizations of their hand-crafted features, i.e., Histogram of Oriented Gradients (HOG) features and $a^{\ast }$ channel of $L^{\ast }{a}^{\ast }{b}^{\ast }$ color space, are employed as three input modalities to train a novel Convolutional Neural Network (CNN). The number of layers, filters, kernel size, learning rate, and optimization technique are carefully selected through an extensive parametric study to minimize the computational cost without compromising accuracy. The system’s performance and robustness are significantly enhanced by jointly deploying the models of these three modalities through ensemble learning. The impact of each modality is optimized based on their impact coefficient determined by grid search. In addition to the comprehensive quantitative assessment, the capabilities of our proposed model and the effectiveness of ensembling over three modalities are evaluated qualitatively using the Grad-CAM visualization model. Experimental results on the test data with additional illumination changes verify the high robustness of our system in dealing with overexposed and underexposed lighting conditions. Achieving a high accuracy ($>99.33\text{\%}$) on six benchmark datasets (i.e., Massey, Static ASL, NUS II, TSL Fingerspelling, BdSL36v1, and PSL) demonstrates that our system notably outperforms the recent state-of-the-art approaches with a minimum number of parameters and high generalization ability over complex datasets. Its promising performance for four different sign languages makes it a feasible system for multi-lingual applications.

In this study, we present a novel graph-based methodology for an accurate classification of cardiac arrhythmia diseases using a single-lead electrocardiogram (ECG). The proposed approach employs the visibility graph technique to generate graphs from time signals. Subsequently, informative features are extracted from each graph and then fed into classifiers to match the input ECG signal with the appropriate target arrhythmia class. The six target classes in this study are normal (N), left bundle branch block (LBBB), right bundle branch block (RBBB), premature ventricular contraction (PVC), atrial premature contraction (A), and fusion (F) beats. Three classification models were explored, including graph convolutional neural network (GCN), multi-layer perceptron (MLP), and random forest (RF). ECG recordings from the MIT-BIH arrhythmia database were utilized to train and evaluate these classifiers. The results indicate that the multi-layer perceptron model attains the highest performance, showcasing an average accuracy of 99.02%. Following closely, the random forest achieves a strong performance as well, with an accuracy of 98.94% while providing critical intuitions.

A collapsed lung, also known as a pneumothorax, is a medical condition characterized by the presence of air in the chest cavity between the lung and chest wall. A chest radiograph is commonly used to diagnose pneumothorax; however, manual segmentation of the pneumothorax region can be difficult to achieve due to its complicated appearance and the variable quality of the image. To address this, we introduce a two-phase deep learning framework designed to enhance the accuracy of lung and pneumothorax segmentation from chest radiographs. Initially, a U-Net model with a ResNet34 backbone, trained on the Shenzhen and Montgomery datasets, is utilized to achieve precise lung region segmentation. Subsequently, for pneumothorax segmentation, we propose the PTXSeg-Net—a convolutional neural network model trained on the SIIM-ACR pneumothorax dataset. The PTXSeg-Net is an enhancement of the U-Net architecture, modified to incorporate attention gates and residual blocks to refine learning capabilities, further strengthened by deep supervision, allowing for more nuanced gradient utilization across all network layers. We employ transfer learning by pre-training an autoencoder to extract robust chest X-ray representations. Data refinement techniques are applied to the SIIM-ACR dataset to further improve training outcomes. Our results indicate that PTXSeg-Net outperforms other models in pneumothorax segmentation, achieving the highest Dice score of 0.9124 and Jaccard index of 0.8894 on the refined dataset with autoencoder pre-training. Moreover, leveraging the predicted lung and pneumothorax segmentation masks from the two-phase framework, we propose a quantification algorithm for estimating the pneumothorax size ratio. Its validity has been confirmed through expert assessments by a radiologist and a surgeon on a test set comprising 495 images. The high acceptance rates, averaging 96.97 %, demonstrate substantial agreement between the proposed method and expert clinical assessments. The implications of these results are significant for clinical practice, offering a deep learning technology for more accurate and efficient pneumothorax identification and quantification. This improvement facilitates the timely determination of required management and treatment strategies, potentially leading to enhancements in patient outcomes.

A Cluster-Based Wireless Sensor Network (CBWSN) is a system designed to remotely control and monitor specific events or phenomena in areas such as smart grids, intelligent healthcare, circular economies in smart cities, and underwater surveillance. The wide range of applications of technology in almost every field of human activity exposes it to various security threats from cybercriminals. One of the pressing concerns that requires immediate attention is the risk of security breaches, such as intrusions in wireless sensor network traffic. Poor detection of denial-of-service (DoS) attacks, such as Grayhole, Blackhole, Flooding, and Scheduling attacks, can deplete the energy of sensor nodes. This can cause certain sensor nodes to fail, leading to a degradation in network coverage or lifetime. The detection of such attacks has resulted in significant computational complexity in the related works. As new threats arise, security attacks get more sophisticated, focusing on the target system's vulnerabilities. This paper proposed the development of Cluster-Based Wireless Sensor Network and Variable Selection Ensemble Machine Learning Algorithms (CBWSN_VSEMLA) as a security threats detection system framework for DoS attack detection. The CBWSN model is designed using a Fuzzy C-Means (FCM) clustering technique, whereas VSEMLA is a detection system comprised of Principal Component Analysis (PCA) for feature selection and various ensemble machine learning algorithms (Bagging, LogitBoost, and RandomForest) for the detection of grayhole attacks, blackhole attacks, flooding attacks, and scheduling attacks. The experimental results of the model performance and complexity comparison for DoS attack evaluation using the WSN-DS dataset show that the PCA_RandomForest IDS model outperforms with 99.999 % accuracy, followed by the PCA_Bagging IDS model with 99.78 % accuracy and the PCA_LogitBoost model with 98.88 % accuracy. However, the PCA_RandomForest model has a high computational complexity, taking 231.64 s to train, followed by the PCA_LogitBoost model, which takes 57.44 s to train, and the PCA_Bagging model, which takes 0.91 s to train to be the best in terms of model computational complexity. Thus, the models surpassed all baseline models in terms of model detection accuracy on flooding, scheduling, grayhole, and blackhole attacks.

This paper is about reducing the power consumption of Cyber–Physical Energy Systems (CPESs) composed of many loads through the usage of a scheduling technique inspired by the real-time-computing domain; the electric-load coordination guarantees a more efficient operation of the entire system by avoiding unnecessary concurrent activation of loads and thus limiting the peak load. We represent the power loads themselves as “physical” components and the computing devices that coordinate them as “cyber” components and formally derive the relationship between the operations in cyber and physical domains as the interplay between the schedules enforced on each component: indeed, the schedule of the loads – generated by combining a two-dimensional bin-packing and an optimal multi-processor real-time scheduling algorithm – influences the timing of the processing tasks that are dedicated to the activation/deactivation of loads themselves. We also consider non-schedulable loads by introducing a policy to cope with the presence of such loads. Numerical simulations and experiments confirm the good performance of the proposed peak load reduction method. The usage of real-time scheduling in this context provides inherent resource optimisation by limiting the number of concurrent loads that are active at the same time, thus directly reducing the overall peak load; moreover, thanks to the limited computational complexity of the algorithms, it scales to large systems, overcoming the scalability issues of common optimisation methods. Numerical simulations and experiments confirm the good performance of the proposed peak load reduction method.

Accounting investing analysis has been expanding at a steady clip, and some of the findings suggest that investors' restricted reasoning and self-psychological sentiments may not always lead them to make their negative emotions influence their financial decisions, leading to a loss. Digital multimedia fusion and display are made possible, and numerous terminals may now communicate with one another in a seamless, real-time manner thanks to the growth of e-commerce and multimedia. This paper proposes a Multimedia Analysis (MA) feature fusion method for understanding the psychological emotions associated with accounting investments in the online retail environment, which can then be used to guide the development of an investment strategy that is both appropriate and successful for the target demographic. The primary goal of this study is to show how multimedia information retrieval tasks may benefit from combining text pre-filtering with image sorting. For this investigation; they used information from the reliable China Stock Market and Accounting Research (CSMAR) Database. The information fusion technology that supports this paper's investigation is used to dissect experiment outcomes, examine issues with the emotional effect of financial investment clients, and assess the paper's intended study topic. In experiments, we found a 97% accuracy rate in terms of accuracy.

A proactive approach is essential to optimize wind turbine maintenance and minimize downtime. By utilizing advanced data analysis techniques on the existing Supervisory Control and Data Acquisition (SCADA) system data, valuable insights can be gained into wind turbine performance without incurring high costs. This allows for early fault detection and predictive maintenance, ensuring that unscheduled or reactive maintenance is minimized and revenue loss is mitigated. In this study, data from a wind turbine SCADA system in the southeast of Ireland were collected, preprocessed, and analyzed using statistical and visualization techniques to uncover hidden patterns related to five fault types within the system. The paper introduces a conditional function designed to test two given scenarios. The first scenario employs a two-tier approach involving fault detection followed by fault identification. Initially, faulty samples are detected in the first tier and then passed to the second tier, which is trained to diagnose the specific fault type for each sample. In contrast, the second scenario involves a simpler solution referred to as naive, which treats fault types and normal cases together in the same dataset and trains a model to distinguish between normal samples and those related to specific fault types. Machine learning models, particularly robust classifiers, were tested in both scenarios. Thirteen classifiers were included, ranging from tree-based to traditional classifiers, neural networks, and ensemble learners. Additionally, an averaging feature importance technique was employed to select the most impactful features on the model decisions as a starting point. A comparison of the results reveals that the proposed two-tier approach is more accurate and less time-consuming, achieving 95% accuracy in separating faulty from normal samples and approximately 91% in diagnosing each fault type. Furthermore, ensemble learners, particularly bagging and stacking, demonstrated superior fault detection and identification performance. The performance of the classifiers was validated using t-SNE and explainable AI techniques, confirming that the impactful features align with the findings and that the proposed two-tier solution outperforms the naive solution. These results strongly indicate that the proposed solution is accurate, independent, and less complex compared to existing solutions.

This research aims to detect different types of Arabic offensive language in twitter. It uses a multiclass classification system in which each tweet is categorized into one or more of the offensive language types based on the used word(s). In this study, five types are classified, which are: bullying, insult, racism, obscene, and non-offensive. To classify the abusive language, a cascaded model consisting of Bidirectional Encoder Representation of Transformers (BERT) models (AraBERT, ArabicBERT, XLMRoBERTa, GigaBERT, MBERT, and QARiB), deep learning models (1D-CNN, BiLSTM), and Radial Basis Function (RBF) is presented in this work. In addition, various types of machine learning models are utilized. The dataset is collected from twitter in which each class has the same number of tweets (balanced dataset). Each tweet is assigned to one or more of the selected offensive language types to build multiclass and multilabel systems. In addition, a binary dataset is constructed by assigning the tweets to offensive or non-offensive classes. The highest results are obtained from implementing the cascaded model started by ArabicBERT followed by BiLSTM and RBF with an accuracy, precision, recall, and F1-score of 98.4%, 98.2%,92.8%, and 98.4%, respectively. RBF records the highest results among the utilized traditional classifiers with an accuracy, precision, recall, and F1-score of 60% for each measurement individually, while KNN records the lowest results obtaining 45%, 46%, 45%, and 43% in terms of accuracy, precision, recall, and F1-score, respectively.

Machine learning applied to regression models offers powerful mathematical tools for predicting responses based on one or more predictor variables. This paper extends the concept of multiple linear regression by implementing a learning system and incorporating both fuzzy predictors and fuzzy responses. To estimate the unknown parameters of this soft regression model, the approach involves minimizing the absolute distance between two lines under three constraints related to the absolute error distance between observed data and their respective predicted lines. A thorough comparative analysis is conducted, showcasing the practical applicability and superiority of the proposed soft multiple linear regression model. The effectiveness of the model is demonstrated through a comprehensive examination involving simulation studies and real-life application examples.