Search

The semiconductor industry is the core industry of the information age. As a key link in the semiconductor industry, wafer fabrication plays a key role in its development. In the testing stage of the wafer, each die of the wafer is detected and marked, and a wafer map with a certain spatial pattern can be formed. The analysis and classification of these spatial patterns can identify the cause of wafer defects, thereby improving production yield. However, as wafer size increases, line widths become smaller, etc., the probability of a mixed defect mode wafer pattern increases. Moreover, the mixed defect mode wafer map is more difficult to identify and classify than the single defect mode wafer map. Therefore, this paper proposes an improved deep convolutional neural network (DCNN) structure model for the recognition and classification of mixed defect pattern wafer maps. From the perspective of increasing the width of the DCNN, the improved network structure can avoid problems such as over-fitting and limited extraction of features due to the continuous deepening of the DCNN. The network is called Multi-Path DCNN (MP-DCNN) structure. The experimental results show that the proposed Multi-Path DCNN structure has better performance and higher classification accuracy than existing methods. [ABSTRACT FROM AUTHOR]

Sentiment analysis extracts valuable insights from textual sources using computation, textual or systematic analysis, and natural language processing. It identifies and measures the attitudes, beliefs, and emotional states individuals express through text data. Recent research on sentiment analysis has largely focused on the English language; therefore, low-resource languages are getting much less attention. Conducting sentiment analysis of low-resource languages is difficult because large datasets and related repositories are unavailable. This paper creates a new dataset for low-resource language (Urdu) to address this issue. The dataset, namely LUCSA-23, consists of more than 65,000 user reviews from various genres, including food, sports, showbiz, apps, and political reviews from developing countries, i.e., Pakistan. Urdu domain experts further annotate the created dataset. This paper proposes an Urdu sentiment analysis approach leveraging the transformer model, i.e., XLM-R and GPT-2. It preprocesses the Urdu text input, generates BERT embeddings, and passes them to the proposed classifier as input for sentiment classification. The proposed classifier is compared with machine/deep/embedded classifiers to evaluate its performance. The findings show that the proposed classifiers outperform existing state-of-the-art approaches with an accuracy of 95%.

Non-technical losses (NTL) pose multiple challenges across distribution grids. This paper introduces a comprehensive framework combining data-driven methods to detect and categorize fraud due to meter tampering or direct connections while identifying potential culprits. The hybrid methodology utilizes grid and consumer-related data to obtain NTL curves through an energy balance approach, yielding indicators such as magnitude, duration, and other features. A Random Forest classifier trained with real historical cases of NTL achieves a weighted F1 score of 0.859, effectively labelling fraud types. Additionally, an unsupervised detection model, integrating clustering and correlation methods, enables accurate identification of tampered meters. The paper introduces two adjustable parameters enabling utilities to fine-tune meter tampering detection strategies based on economic considerations. The results demonstrate that true positives can be increased at the expense of increasing false positives. Accurate fraud identification is achieved using the Fuzzy C-Means algorithm, with an F1 score of 0.9. The algorithm is tested on grids with distributed generation, with a decrease of 10% on the predictive performance when half of the users are prosumers, demonstrating the methodology’s promising performance in real-world scenarios.

Identifying lung sound signal patterns is essential for detecting and monitoring respiratory diseases. Existing approaches for analyzing respiratory sounds need domain specialists. Therefore, an accurate and automated lung sound classification tool is required. In this paper, we have developed an automatic diagnostic system to classify these signals. It can support healthcare systems in low-resource environments with limited resources and a shortage of qualified medical professionals. This paper presents an eigenvectors-based data augmentation method to enhance the detection rate of automatic diagnostic systems. This proposed method provides noise-free data samples with the principal components that capture the most significant variations in the data. In the classification process, various machine learning-based classifiers are employed along with spectrogram-based features.

Chromosome karyotyping is an important yet labor-intensive procedure for diagnosing genetic diseases. Automating such a procedure drastically reduces the manual work of cytologists and increases congenital disease diagnosing precision. Researchers have contributed to chromosome segmentation and classification for decades. However, very few studies integrate the two tasks as a unified, fully automatic procedure or achieved a promising performance. This paper addresses the gap by presenting: 1) A novel chromosome segmentation module named ChrRender, with the idea of rendering the chromosome instances by combining rich global features from the backbone and coarse mask prediction from Mask R-CNN; 2) A devised chromosome classification module named ChrNet4 that pays more attention to channel-wise dependencies from aggregated informative features and calibrating the channel interdependence; 3) An integrated Render-Attention-Architecture to accomplish fully automatic karyotyping with segmentation and classification modules; 4) A strategy for eliminating differences between training data and segmentation output data to be classified. These proposed methods are implemented in three ways on the public Q-band BioImLab dataset and a G-band private dataset. The results indicate promising performance: 1) on the joint karyotyping task, which predicts a karyotype image by first segmenting an original microscopical image, then classifying each segmentation output with a precision of 89.75% and 94.22% on the BioImLab and private dataset, respectively; 2) On the separate task with two datasets, ChrRender obtained AP50 of 96.652% and 96.809% for segmentation, ChrNet4 achieved 95.24% and 94.07% for classification, respectively. The COCO format annotation files of BioImLab used in this paper are available at https://github.com/Alex17swim/BioImLab The study introduces an integrated workflow to predict a karyotyping image from a Microscopical Chromosome Image. With state-of-the-art performance on a public dataset, the proposed Render-Attention-Architecture has accomplished fully automatic chromosome karyotyping.

In the real world, there are only a small amount of data with labels. To make full use of the potential structural information of unlabeled data to train a better classifier, researchers have proposed many semi-supervised learning algorithms. Among these algorithms, self-training is one of the most widely used semi-supervised learning frameworks due to its simplicity. How to select high-confidence samples is a crucial step for self-training. If the misclassified samples are selected as high-confidence samples, this error will be amplified in the iterative process, which affects the performance of the final classifier. To alleviate the impact of this problem, this paper proposes a self-training algorithm with block-similar neighbor editing (STBSNE). STBSNE calculates the distance between samples by the block-based dissimilarity measure, which improves the classification performance on high-dimensional data sets. STBSNE defines the block-estimated neighbor relationship, builds the block-estimated neighbor relationship graph, and proposes the block estimated neighbor editing method to identify outliers and noise points, and edits them to improve the quality of the high-confidence sample selected. Experimental results on 16 benchmark data sets verify the superior performance of the proposed STBSNE compared with seven state-of-the-art algorithms.

Nowadays, we witness rapid technological advancements in online communication platforms, with increasing volumes of people using a vast range of communication solutions. The fast flow of information and the enormous number of users opens the door to the publication of non-truthful news, which has the potential to reach many people. Disseminating this news through low- or no-cost channels resulted in a flood of fake news that is difficult to detect by humans. Social media networks are one of these channels that are used to quickly spread this fake news by manipulating it in ways that influence readers in many aspects. That influence appears in a recent example amid the COVID-19 pandemic and various political events such as the recent US presidential elections. Given how this phenomenon impacts society, it is crucial to understand it well and study mechanisms that allow its timely detection. Deep learning (DL) has proven its potential for multiple complex tasks in the last few years with outstanding results. In particular, multiple specialized solutions have been put forward for natural language processing (NLP) tasks. In this paper, we systematically review existing fake news detection (FND) strategies that use DL techniques. We systematically surveyed the existing research articles by investigating the DL algorithms used in the detection process. Our focus then shifts to the datasets utilized in previous research and the effectiveness of the different DL solutions. Special attention was given to the application of strategies for transfer learning and dealing with the class imbalance problem. The effect of these solutions on the detection accuracy is also discussed. Finally, our survey provides an overview of key challenges that remain unsolved in the context of FND.

Melanocytic nevi (MN), which are the most prevalent benign skin tumors, can be classified into three subtypes based on the depth of the nevus cell nests: junctional, compound, and intradermal nevi. Among these subtypes, intradermal nevi pose a lower risk of malignancy, whereas junctional nevi carry a higher risk of malignancy. To facilitate early stage diagnosis, mitigate unnecessary biopsies, and reduce medical costs, this paper presents the first study on the classification of benign melanocytic nevi. We propose an algorithm that utilizes depth-relevant, channel-specific, and channel-shared information from dermoscopy images. This approach draws inspiration from the physical principles underlying dermoscopy imaging and the physiological essence of nevi. We employed CNN and transformer blocks to capture local and long-range contextual information. In addition, our physiology-inspired module focuses on learning depth-related features from the color channels. To assess the performance of our proposed method, we curated a dataset of dermoscopic images of MN and compared it with two human experts as well as classic CNN and Transformer baseline models. Our results showed that the proposed automated method achieved an average accuracy of 70.23%, surpassing the best-performing baseline by 3.6%, and outperforming human experts by nearly 20%. Furthermore, we conducted an ablation study to confirm the effectiveness of our algorithm and to analyze its clinical significance.

The main goal of this paper is to introduce a Motor Imagery (MI) classification system for electroencephalography (EEG) that is extremely precise. To achieve this goal, we propose using a feature-extracted deep one-dimension (1D) convolutional neural network (CNN) which provides a model that can be further improved through hyperheuristic multi-objective evolutionary search. We can improve the classification performance by training this deep CNN model with feature-extracted data from the Physionet MI dataset. We also present a semi-deep fine-tuning approach that can yield improvements with just four epochs. Our findings using the Physionet MI dataset illustrate that the approach we suggest surpasses most contemporary techniques used for classifying EEG signals. Our system is computationally efficient and can be trained using reliable EEG data for individual patients, allowing for accurate classification of their EEG records. Because of its straightforward and parameter-independent characteristics, our system is versatile and can be utilized with any EEG dataset.

Predictive analytics is concerned with making predictions of future outcomes based on past data using data statistics, machine learning, dynamic models and filtering algorithms. This paper provides a review of the theory of predictive analytics from an optimization perspective, including some new developments in convex optimization and quadratic neural networks. Several examples from many different areas show how the theory can be applied to analyze data and provide useful information to decision makers.

The signal-to-noise ratio (SNR) is an important metric for measuring signal quality and its estimation has received widespread attention in various application scenarios. In this paper, we propose an SNR estimation framework based on deep learning classification. Power spectrum input is proposed to reduce the computational complexity. We also propose an SNR estimation method based on deep learning regression to overcome the inevitable estimation error problem of classification-based methods in dealing with signals with SNR not within the training label set. We conduct a large number of simulation experiments considering various scenarios. Results show that the proposed methods have better estimation accuracy than two existing deep learning-based SNR estimation methods in different noises and multipath channels. Furthermore, the proposed methods only need to be trained under one modulation signals to adapt to SNR estimation of other modulation signals, with superior transfer performance. Finally, the method using the average periodogram as input has stronger adaptability in few-shot scenario and requires lower computational complexity compared to the method with in-phase and quadrature (IQ) input.

In this paper, we explore the applicability of selected machine learning models to classify incoming flows as elephants or mice on the first packet, using Internet Protocol (IP) and transport layer headers (5-tuple). We show that traditional metrics such as accuracy or F1-score are inadequate for assessing performance in traffic engineering (TE) and quality of service (QoS) applications unless compared at the same traffic coverage. Among the classifiers analyzed, Histogram-based Gradient Boosting with octets-transformed input data provides the best performance, reducing flow operations by a factor of 36.49 and the average number of flow table entries by 16.35, while covering 80% of the traffic.

Cervical spine diseases, encompassing conditions like spondylolisthesis, disc degeneration, and cervical spinal stenosis, stand as significant contributors to global disability. Precise classification of these conditions is paramount for effective medical diagnosis. This paper introduces an innovative methodology aimed at addressing the limitations of traditional convolutional neural networks (CNNs) and pretrained models in this domain. We propose a novel approach dubbed Inv-AlxVGGNets, which leverages concatenated pretrained architectures AlexNet and VGG, augmented by involutional neural networks and residual layers. Unlike conventional CNNs that are location-specific and channel-agnostic, involutional neural networks offer enhanced Adaptability to diverse visual patterns in medical images. Focusing on a four-class cervical spine disease classification task utilizing MRI images, our study evaluates the performance of Inv-AlxVGGNets (AlexNet with INN and VGG with residual layer models) as well as machine learning algorithms. Our findings demonstrate superior performance in terms of accuracy, precision, recall, and AUC ROC values. Notably, Inv-AlxVGGNets achieves an impressive 98.73% accuracy on the testing set and 99.78% on the training set, underscoring its potential for precise cervical spinal disease classification. In a comparative analysis, we highlight that conventional CNNs entail over 133 million parameters, whereas Inv-AlxVGGNets require less than 8 million parameters, rendering them more efficient and resource-friendly. This reduced parameter count is particularly advantageous in resource-constrained scenarios, where computational resources and datasets may be limited. The promising results position Inv-AlxVGGNets as a valuable tool for precise cervical spine disease classification, offering implications for enhancing patient care in resource-constrained settings.

This paper evaluates the performance of six different machine learning (ML) algorithms for classifying power quality disturbances (PQDs), with statistical features extracted using discrete wavelet transform (DWT) as feature input. The statistical features have been extracted from coefficients of multi-resolution analysis (MRA) using four different mother wavelets: Daubechies 4 (‘db4’), ‘haar’, Discrete Meyer (‘dmey’), Coiflets 4 (‘coif4’). The performance analysis has been carried out with 5,500 synthetic signals pertaining to eleven different PQDs generated in accordance with IEEE 1159-2019. Moreover, the performance of the classifiers trained with synthetic signals has been investigated under the influence of unseen noisy signals, hardware PQD signals obtained from the experimental setup, and real PQD events. The analysis indicates that the performance of the extra tree (ET) classifier with the features extracted using ‘haar’ as a mother wavelet is superior and robust in comparison to other classifiers, viz k-nearest neighbor (kNN), random forest (RF), decision tree (DT), logistic regression model (LRM), and gaussian naïve bayes (GNB) with features extracted using different mother wavelets. Furthermore, the ‘haar-ET’ based technique demonstrated remarkable performance in classifying PQDs, showing strong generalization to both unseen hardware and noisy signals, and achieving consistent results when tested with real PQD events.

Bruxism is a sleep disorder which is manifested by unintentional grinding and clenching of teeth during sleep. An automated sleep bruxism recognition system using single channel EEG data is proposed in this paper which is based on Inverse Discrete Wavelet Transformed Reconstructed Band Limited (IDWT-RBL) signals. These band limited EEG signals are used for extracting various features. Instead of using handcrafted features, feature reduction is done by ranking using statistical test scoring combined with classifier testing. This technique finds optimal features, reduces model complexity, lowers computational burden and increases model interpretability. Abundant features from time, frequency and statistical domains are used primarily so that no significant feature is ignored. Choosing a good subset of features from a larger set is a challenge for data with low sample size. To meet this challenge, a Group wise Feature Ranking (GFR) technique is introduced to reduce feature dimension. After statistical ranking and group wise averaging the scores, most significant groups of features are chosen. The proposed scheme is validated on a publicly available dataset. This process is examined for both unlabeled and labeled sleep stage. For segments with unlabeled sleep stage, cubic Support Vector Machine (SVM) performed best for F3C3 channel using 6 features with an accuracy of 97.83%. For segments with labeled sleep stage, F3C3 and REM sleep stage using 10 features performed best with 98.39% accuracy. The accuracy of proposed method is superior to most recent bruxism detection techniques. Finally, the GFR technique is applied to detect sleep disordered breathing (SDB). The outstanding performance to detect SDB clearly demonstrates the versatility of the proposed GFR technique to solve binary classification problem using EEG signals. Moreover, the introduced GFR technique enhances confidence and pellucidity of the system as it is more explainable.

This study explores the contemporary landscape of integrating numerical algorithms, artificial intelligence (AI), and expert methodologies within the domain of nephrology. Focusing on automation and decision support, we scrutinize the impact of numerical algorithms on the precise evaluation of kidney function. Furthermore, we delve into the transformative potential of artificial intelligence, particularly in the realms of machine learning and deep learning applications, elucidating its role in early disease detection and the formulation of personalized treatment strategies. The synergy between computer-based tools and expert-driven approaches is examined, underscoring their collaborative role in enhancing the accuracy and dependability of diagnoses. Additionally, we address the ethical considerations and challenges associated with the incorporation of automation in nephrology. The paper provides a concise overview of in-depth analysis while illustrating the promising prospects of these innovative methodologies for reshaping nephrology research and patient care.

Automatic and accurate classification of thyroid nodules is of great significance to doctors for clinical diagnosis and subsequent treatment recommendations. Since there are no obvious features between benign and malignant nodules, enlarging or reducing the image will result in blurred edges and image distortion, thus limiting the accuracy of clinical diagnosis. Furthermore, the prevalence of sample class imbalance in medical images poses significant challenges to applying convolutional neural networks in thyroid nodule classification methods. This paper proposes a network Tnc-Net for thyroid nodule classification. The network backbone can adapt to the problem of small data volume, capture global features with the help of simple channel attention, and effectively extract image information. The branch network supplements the feature extraction from the backbone network, and the information extracted from the backbone and branch networks is effectively utilized through the fusion module. In addition, this article designs training strategies suitable for this network to deal with category imbalance, improve model classification performance, and make classification results more clinically referenceable. The method test accuracy is 0.902, which exceeds other classic deep learning models in classification. This result demonstrates the effectiveness of our method in achieving automatic classification of thyroid nodules.

Goal: The purpose of this paper is to recognize autism spectrum disorders (ASD) using graph attention network. Methods: we propose a node features graph attention network (NF-GAT) for learning functional connectivity (FC) features to achieve ASD diagnosis. Firstly, node features are modelled based on functional magnetic resonance imaging (fMRI) data, with each subject modelled as a graph. Next, we use the graph attention layer to learn the node features and gets the node information of different nodes for ASD classification. Results: Compared with other models, the NF-GAT has significant advantages in terms of classification results. Conclusions: NF-GAT can be effectively used for ASD classification.

Parkinson’s disease is a progressive neurological condition that affects motor abilities. Common symptoms include tremors, muscle stiffness, and difficulty with coordinated movements. A variety of efforts are underway to address these issues and improve diagnostic precision in Parkinson’s disease. This paper employs well-known machine-learning techniques to improve diagnostic accuracy. A variety of individual and ensemble AI models have been proposed, including Random Forest, Decision Tree, Logistic Regression, Gradient Boosting, Support Vector Machine, Stacking, and Bagging Ensemble classifiers. Three scenarios are applied to two standard benchmark datasets. The best performance is achieved when the Stacking Ensemble classifier is utilized, where the Support Vector Machine and Gradient Boosting are engaged for extracting features and Logistic Regression for classifying Parkinson’s disease. The Stacking Ensemble classifier reaches 94.87% accuracy and 90.00% AUC for the first dataset, while for the second dataset, 96.18% accuracy and 96.27% AUC are recorded. The final results demonstrate the importance of the suggested framework, which can help to improve the overall diagnosis outcomes.

As one of the popular and effective supervised classification methods, linear regression is extensively used in image classification. However, the zero-one labeling matrix is too strict to be conducive to linear regression methods for learning labeling information. In addition, the linear regression focuses only on the fit of the input features to the corresponding output labels and ignores the distinctiveness between the samples. To address these two issues, this paper proposes a new method, namely, discriminative regression with latent label learning, for image classification. In contrast to the other methods, the proposed method learns labeling information in the latent label space instead of the input zero-one labeling space, doing so has the advantage that the proposed method can learn the labeling information in the data more flexibly. To guide the transform matrix to learn the discriminative information in the data, a regularization term with the idea of shortening the distance between samples within a class and lengthening the distance between samples between classes is integrated into the objective function of the proposed method. To obtain the solution of the proposed model, an iterative optimization algorithm is developed. Comprehensive experiments show that the classification performance of the proposed method outperforms the current state-of-the-art methods and deep learning methods on public image datasets with small sample sizes.

Educational Data Mining (EDM) is used to ameliorate the teaching and learning process by analyzing and classifying data that can be applied to predict the students’ academic performance, and students’ dropout rate, as well as instructors’ performance. The prediction of student performance is complicated by the vast and diverse range of variables from academic records to behavioral and health metrics. In this paper, we have introduced a new Adaptive Feature Selection Algorithm (AFSA) by amalgamating an ensemble approach for initial feature ranking with normalized mean ranking from five distinct methods to enhance robustness. The proposed method iteratively selects the best features by adjusting its threshold based on each feature’s rank to ensure significant contributions to model accuracy and also effectively reduces dataset complexity. We have tested the performance of the proposed feature selection algorithm using five machine learning classifiers: Logistic Regression (LR), K-Nearest Neighbour (KNN), Support Vector Machine (SVM), Naïve Bayes (NB) classifier, and Decision Tree (DT) classifier on four student performance datasets. The experimental results highlight the proposed method significantly decreases feature count by an average feature reduction factor of 5.7, significantly streamlining datasets while maintaining competitive cross-validation accuracy, marking it as a valuable tool in the field of educational data analytics.

Malaria is a mosquito-borne, life-threatening, and contagious disease that has caused thousands of fatalities in recent years. Due to inadequate detection, the inexperience of laboratory personnel, and lack of advanced point-of-care equipment, the malaria-induced mortality rate is increasing. In addition to the traditional detection mechanisms, researchers have recently been investigating microscopic malaria-infected Red Blood Cells (RBC) image analysis based on deep learning models to detect malaria parasites as a general-purpose point-of-care solution. In this paper, we develop several hybrid data-driven models by combining a convolutional neural network (CNN) to extract the relevant features and two cascaded recurrent neural networks (RNNs) classifiers. Gated recurrent unit (GRU), long short-term memory (LSTM), and Bi-directional LSTM (BiLSTM) are considered candidate RNN classifiers. The models are compared in terms of accuracy, type-I & II error rates, and inference & training computation time. Experimental results demonstrate that the CNN-LSTM-BiLSTM model outperforms the other models with a significantly higher accuracy (96.20%), less type-I error rate (2.23%), and fewer combined type-I and type-II errors (3.80%). Also, we consider the model computation time (both training time per epoch and inference time per step)‘ as an important metric for emerging distributed learning paradigms where point-of-care devices with IoT (Internet of Things) capability can jointly contribute to global model accuracy. Thus, our findings demonstrate the practicality of cascading classifiers in resource-constrained point-of-care devices.

Invasive Coronary Angiography (ICA) images are considered the gold standard for assessing the state of the coronary arteries. Deep learning classification methods are widely used and well-developed in different areas where medical imaging evaluation has an essential impact due to the development of computer-aided diagnosis systems that can support physicians in their clinical procedures. In this paper, a new performance analysis of deep learning methods for binary ICA classification with different lesion degrees is reported. To reach this goal, an annotated dataset of ICA images containing the ground truth, the location of lesions, and seven possible severity degrees ranging between 0% and 100% was employed. The ICA images were divided into “lesion” or “non-lesion” patches. We aim to study how binary classification performance is affected by the different lesion degrees considered in the positive class. Therefore, five Convolutional Neural Network architectures – DenseNet-201, MobileNet-V2, NasNet-Mobile, ResNet-18, and ResNet-50 – were trained with different input images where different lesion degree ranges were gradually incorporated until considering the seven lesion degrees. Besides, four types of experiments with and without data augmentation were designed, whose F-measure and Area Under Curve (AUC) were computed. Reported results achieved an F-measure and AUC of 92.7% and 98.1%, respectively. However, lesion classification is highly affected by the degree of the lesion intended to be classified, with 15% less accuracy when < 99% lesion patches are present.

The paper presents a novel recognition framework for incomplete noninvasive Electroencephalography (EEG) signals relying on the recent advances in tensor algebra, named as An Innovative Tensor Algebra-based Detection Framework for Incomplete Noninvasive Electroencephalography (iTa-DFiE). iTa-DFiE is motivated to improve the diagnostic performance by tackling the major problems shared by a variety of noninvasive EEG-based Brain-Computer Interfaces (BCIs) application is tensorial structured time series with occlusions. The aforementioned challenge setting is solved on two major thrusts, including: 1) tensor completion: discovering hidden patterns and learning their evolving trends to offer missing values imputation via improvement of standard Kalman Filter approach and 2) tensor decomposition: extracting essential hidden information from multi aspects data tensor via extending the most well-known tensor factorization Tucker. The effectiveness and efficiency of the proposed tensor-based framework is proved via successfully improving the pattern classification results on two real-world noninvasive EEG-based motor imagery BCI with diverse corrupted data scenarios, especially in occurrence of consecutive missing observations. Strikingly, iTa-DFiE also outperforms the conventional matrices-based methods and the state-of-the-art tensor techniques in terms of missing reconstructions and feature extraction as well.

Online transient stability assessment (TSA) is essential for the reliable operation of power systems. The increasing deployment of phasor measurement units (PMUs) across power systems provides a wealth of fast, accurate, and detailed transient data, offering significant opportunities to enhance online TSA. Unlike conventional data-driven methods that require large volumes of transient PMU data for accurate TSA, this paper develops a new TSA method that requires significantly less data. This data reduction is enabled by generative and adversarial networks (GAN), which predict voltage time-series data following a transient event, thereby minimizing the need for extensive data. A classifier embedded in the generative network deploys the predicted data to determine the stability of the system. The developed method preserves the temporal correlations in the multivariate time series data. Hence, compared to the state-of-the-art methods, it is more accurate using only one sample of the measured PMU data and has a shorter response time.

This paper presents an ensemble method for binary classification, where each base model is based on an extended neighbourhood rule (ExNRule). The ExNRule identifies the neighbours of an unseen observation in a stepwise manner. This rule first selects the sample point closest to the experimental observation and then selects the second observation nearest to the previously chosen one. To find the required data points in the neighbourhood, this search is repeated up to k steps. The test sample point is predicted using majority voting in the class labels of the k chosen neighbours. In the proposed method, a large number of ExNRule based models are constructed on randomly projected bootstrap samples. The error rates of these models are computed using out-of-bag data points. The models are then ranked according to their out-of-bag errors, and a proportion of the most accurate models are selected. The final ensemble is constructed by combining the selected models. The proposed method is compared with other classical procedures on 15 benchmark datasets in terms of classification accuracy, Kohen’s kappa and Brier score (BS) as performance metrics. Boxplots of the results are also constructed. The proposed ensemble is outperforming the existing methods on almost all the benchmark datasets. For further evaluation, the proposed method is compared with other kNN based classifiers on 3 datasets using different k values. Furthermore, the performance of the proposed method is also evaluated using simulated data under different scenarios.

AutoML offers significant benefits in solving real-life problems because it accelerates the development of machine learning models. In contexts involving real scenarios like analyzing companies’ environmental, social and governance (ESG), where the dataset presents some challenges, AutoML is anticipated as a promising solution to address these complexities. Although researchers have shown significant interest in exploring Genetic Programming (GP) in AutoML for handling complex datasets, a critical issue that remains unresolved is the comprehensive understanding of GP hyper-parameters that influence machine learning performance. While GP-based AutoML excels in automating many aspects of the modelling, there has been a scarcity of research that provides insight into the significance of individual features and GP population size within the models of GP-based AutoML. This paper presents a comprehensive analysis of the models’ performance evaluation from multiple facets, including feature selection, GP population sizes, and different machine learning algorithms. Furthermore, this study provides insights into the association between Pearson correlations, machine learning performance, and the importance of machine learning features. The findings demonstrate that incorporating all the determinants as features in GP-based AutoML or relying solely on firm characteristics led to superior performance with an excellent trade-off between True Positive Rate and False Positive Rate. Thus, higher accuracy results exceeding 0.9 of Area Under the Curve (AUC) are presented by the proposed models. The novelty of this study lies in its empirical evaluation of different approaches to GP-based AutoML implementation. These findings provide alternative solutions for business investors to identify companies with strong sustainability practices.

The widespread use of artificial intelligence (AI) in more and more real-world applications is accompanied by challenges that are not obvious at first glance. In machine learning, class imbalance, characterized by an imbalance in the frequency of classes, is one key challenge that poses essential problems for many common machine learning algorithms. This challenge led to the development of various countermeasures to tackle class imbalance. Although these countermeasures improve the prediction performance of models, they often jeopardize interpretability for both AI users and AI experts. Especially in sensitive domains where class imbalance is regularly present, for example, medicine, meteorology, or fraud detection, interpretability is of utmost importance. In this paper, we evaluate the effect of class imbalance countermeasures on interpretability with methods of explainable AI (XAI). Our work contributes to a more in-depth understanding of these countermeasures and connects the research fields of class imbalance learning and XAI. Our experimental results suggest that only feature selection and cost-sensitive approaches are the only class imbalance countermeasures that preserve interpretability for both AI users and AI experts. In contrast, resampling and most classification algorithms for imbalance learning are not suitable in settings where knowledge should be derived and where interpretability is a key requirement.

Accents, or changes in how different people speak the same word/sentence in the same language, pose substantial communication issues in most spoken languages. This is a well-known fact, but how does the accent of one language affect learning/speaking another? In this paper, we look at how Arab accents influence the English language. To that end, we built a deep machine-learning system for Arabic accent recognition that was learned from an in-house English speech database of four Arabic accents collected from Jordan, Iraq, Saudi Arabia, and Tunisia. The proposed system employs Mel spectrograms of an English-spoken paragraph to train an LSTM neural network to recognize the accent in each sound signal. Although the collected data was extremely difficult to learn due to the presence of both males and females and fluent speakers in each class, the proposed system could recognize speakers with various accents by up to 79%. This answers the study’s main question, demonstrating that speakers with an Arabic accent have their way of speaking English, which varies by country. As a result, if trained on appropriate and adequate data, the proposed system can also be used to recognize accents in any language.

Sleep disorder classification is crucial in improving human quality of life. Sleep disorders and apnoea can have a significant influence on human health. Sleep-stage classification by experts in the field is an arduous task and is prone to human error. The development of accurate machine learning algorithms (MLAs) for sleep disorder classification requires analysing, monitoring and diagnosing sleep disorders. This paper compares deep learning algorithms and conventional MLAs to classify sleep disorders. This study proposes an optimised method for the Classification of Sleep Disorders and uses the Sleep Health and Lifestyle Dataset publicly available online to evaluate the proposed model. The optimisations were conducted using a genetic algorithm to tune the parameters of different machine learning algorithms. An evaluation and comparison of the proposed algorithm against state-of-the-art machine learning algorithms to classify sleep disorders. The dataset includes 400 rows and 13 columns with various features representing sleep and daily activities. The k-nearest neighbours, support vector machine, decision tree, random forest and artificial neural network (ANN) deep learning algorithms were assessed. The experimental results reveal significant performance differences between the evaluated algorithms. The proposed algorithms obtained a classification accuracy of 83.19%, 92.04%, 88.50%, 91.15% and 92.92%, respectively. The ANN achieved the highest classification accuracy of 92.92%, and its precision, recall and F1-score values on the testing data were 92.01%, 93.80% and 91.93%, respectively. The ANN algorithm that achieved high accuracy than other tested algorithms.

Art objects can evoke certain emotions. Color is a fundamental element of visual art and plays a significant role in how art is perceived. This paper introduces a novel approach to classifying emotions in art using Fuzzy Sets. We employ a fuzzy approach because it aligns well with human judgments’ imprecise and subjective nature. Extensive fuzzy colors (n=120) and a broad emotional spectrum (n=10) allow for a more human-consistent and context-aware exploration of emotions inherent in paintings. First, we introduce the fuzzy color representation model. Then, at the fuzzification stage, we process the Wiki Art Dataset of paintings tagged with emotions, extracting fuzzy dominant colors linked to specific emotions. This results in fuzzy color distributions for ten emotions. Finally, we convert them back to a crisp domain, obtaining a knowledge base of color-emotion associations in primary colors. Our findings reveal strong associations between specific emotions and colors; for instance, gratitude strongly correlates with green, brown, and orange. Other noteworthy associations include brown and anger, orange with shame, yellow with happiness, and gray with fear. Using these associations and Jaccard similarity, we can find the emotions in the arbitrary untagged image. We conducted a 2AFC experiment involving human subjects to evaluate the proposed method. The average hit rate of 0.77 indicates a significant correlation between the method’s predictions and human perception. The proposed method is simple to adapt to art painting retrieval systems. The study contributes to the theoretical understanding of color-emotion associations in art, offering valuable insights for various practical applications besides art, like marketing, design, and psychology.

Security attacks are becoming more sophisticated and common as connected devices rapidly exchange personal, sensitive, and important data. Security solutions are therefore required for Internet of Things (IoT) environments. System administrators receive alerts through an automatic Network Intrusion Detection (NID) system when security breaches occur. An automatic NID can be an effective tool to protect IoT networks against various attacks. It is possible to detect intrusions using a variety of intrusion detection techniques, but the performance and class imbalance in the dataset make this a difficult process. To improve detection rates and decrease false alarms, intrusion detection accuracy must be improved. In this paper, an automatic NID system is proposed leveraging a renowned machine learning model named Random Forest (RF) on the (UNSW-NB15) dataset collected from Kaggle. The experimental results indicate that the proposed model not only has higher accuracy at 90.17% surpassing the baseline approach by 7.34%, but also has precision, recall, and F1 scores up to 90.14%, 90.17, and 90.14%, respectively. Moreover, 98.83% accuracy is achieved with a balanced class dataset by using random resampling techniques to generate synthetic data of minority attacks.

The increased utilization of digital instruments like smartphones, Internet of Things (IoT) sensors, cameras, and microphones has resulted in extensive amounts of big data. Inherent challenges associated with big data include significant data dimensionality, redundancy, and irrelevant information. The main objective of feature selection is eliminating unnecessary features, thereby minimizing time and space requirements. This paper proposes a new Binary Crayfish Optimization Algorithm (BinCOA) for feature selection. The Crayfish Optimization Algorithm (COA) is a new metaheuristic algorithm inspired by the simulation of Crayfish search for food, summer resorts, and competitive habits. The original COA has been augmented with two primary enhancements to improve its performance. The refracted opposition-based learning strategy is a novel enhancement incorporated into the initialization step of the COA algorithm to strengthen the algorithm’s capability for exploitation. The crisscross strategy is added to the original COA, increasing the COA’s convergence accuracy. The algorithm’s performance is assessed by evaluating a set of 30 benchmark datasets. The proposed BinCOA is evaluated in comparison to seven contemporary wrapper feature selection methods. The experimental finding indicates that BinCOA consistently outperforms existing algorithms in classification accuracy, average fitness value, and the number of selected features. Furthermore, the statistical significance of the results is verified by calculating the Wilcoxon rank-sum test.

Degenerative diseases of the lumbar spine, such as spondylolisthesis, disc degeneration, and lumbar spinal stenosis, are major contributors to global disability. Accurate classification of lumbar diseases is crucial for effective medical diagnosis. This paper introduces an innovative methodology for lumbar disease classification, addressing the limitations of traditional convolutional neural networks (CNNs). We propose a novel approach, InVGG, which combines involutional neural networks with the VGG architecture. Unlike traditional CNNs, InVGG utilizes involution kernels that are location-specific and channel-agnostic, enhancing its adaptability to varied visual patterns in medical images. Our study focuses on a four-class lumbar disease classification problem using sagittal T2 MRI images. The evaluation of InVGG is compared with traditional CNNs (VGG model) and machine learning algorithms, demonstrating superior performance in terms of accuracy, precision, recall, and AUC ROC values. InVGG achieves an impressive 96% accuracy on the testing set and 99% on the training set, showcasing its potential for accurate spinal lumbar disease classification. The reduced parameter count of InVGG compared to CNNs (VGG) makes it more resource-efficient, especially in scenarios with limited computational resources and datasets. The promising results position InVGG as a valuable tool for precise lumbar disease classification, with implications for improving patient care in resource-constrained scenarios.

Automation improves the quality of fruits through quick and accurate detection of pest and disease infections, thus contributing to the country’s economic growth and productivity. Although humans can identify the fruit damage caused by pests and diseases, the methods used are inconsistent, time-consuming, and variable. The surface features of fruits typically observed by consumers who seek their health benefits affect their market value. The issue of pest and disease infections further deteriorates fruits’ quality, becoming a mounting stressor on farmers since they reduce the potential revenue from fruit production, processing, and export. This article reviews various studies on detecting and classifying damages in fruits. Specifically, we review articles where state-of-the-art approaches under segmentation, image processing, machine learning, and deep learning have proved effective in developing automated systems that address hurdles associated with manual methods of assessing damage using visual experiences. This survey reviews thirty-two journal and conference papers from the past thirteen years that were found electronically through Google Scholar, Scopus, IEEE, ScienceDirect, and standard online searches. This survey further presents a detailed discussion of previous research done in the past while emphasizing their strengths and limitations as well as outlining potential future research topics. It also reveals that much as the use of automated detection and classification of fruit damage has yielded promising results in the horticulture industry, more research is still needed with systems required to fully automate the detection and classification processes, especially those that are mobile phone-based towards addressing occlusion challenges.

This study presents a novel method to assess the learning effectiveness using Electroencephalography (EEG)-based deep learning model. It is difficult to assess the learning effectiveness of professional courses in cultivating students’ ability objectively by questionnaire or other assessment methods. Research in the field of brain has shown that innovation ability can be reflected from cognitive ability which can be embodied by EEG signal features. Three navigation tasks of increasing cognitive difficulty were designed and a total of 41 subjects participated in the experiment. For the classification and tracking of the subjects’ EEG signals, a convolutional neural network (CNN)-based Multi-Time Scale Spatiotemporal Compound Model (MTSC) is proposed in this paper to extract and classify the features of the subjects’ EEG signals. Furthermore, Spiking neural networks (SNN) -based NeuCube is used to assess the learning effectiveness and demonstrate cognitive processes, acknowledging that NeuCube is an excellent method to display the spatiotemporal differences between spikes emitted by neurons. The results of the classification experiment show that the cognitive training traces of different students in solving three navigational problems can be effectively distinguished. More importantly, new information about navigation is revealed through the analysis of feature vector visualization and model dynamics. This work provides a foundation for future research on cognitive navigation and the training of students’ navigational skills.

The mining and metal processing industries are undergoing a transformation through digitization, with sensors and data analysis playing a crucial role in modernization and increased efficiency. Vibration sensors are particularly important in monitoring production infrastructure in metal processing plants. This paper presents the installation of vibration sensors in an actual industrial environment and the results of spectral vibration data analysis. The study demonstrates that vibration sensors can be installed in challenging environments such as metal processing plants and that analyzing vibration patterns can provide valuable insights into predicting machine failures and different machine states. By utilizing dimensionality reduction and dominant frequency observation, we analyzed vibration data and identified patterns that are indicative of potential machine states and critical events that reduce production throughput. This information can be used to improve maintenance, minimize downtime, and ultimately enhance the production process’s overall efficiency. This study highlights the importance of digitization and data analysis in the mining and metal processing industries, particularly the capability not only to predict critical events before they impact production throughput and take action accordingly but also to identify machine states for legacy equipment and be part of retrofitting strategies.

Coral reefs are among the most biologically diverse and economically valuable ecosystems on Earth, but they are threatened by climate change. Understanding how reefs developed over geological timescales can provide important information about past environmental changes and their impacts on reef systems. Significant effort and capital have been invested in drilling and analyzing reef cores. Recognizing coral and sediment patterns visually from fossil reefs is a laborious task that demands domain expertise. In this paper, we present a machine learning-based framework that utilizes clustering and classification methods to fuse multiple sources of data for the segmentation and annotation of reef cores. The framework produces an annotated image of a reef core with six lithologies identified; massive corals, encrusted corals, coralline algae, microbialite, sand, and silt. We utilize reef cores recovered from Expedition 325 of the International Ocean Discovery Program (IODP) to the Great Barrier Reef. We use reef core image data and physical properties data to segment reef cores. We evaluate the framework using selected clustering and classification models. The results show that Gaussian mixture models can provide accurate segmentation of reef core image data, with a clear visual distinction between two major classes: massive corals and stromatolitic microbialites. Furthermore, we find that the random forest classifier provides the best annotations for the segmented reef core image data with an accuracy of 96%.

Data privacy is essential in the financial sector to protect client’s sensitive information, prevent financial fraud, ensure regulatory compliance, and safeguard intellectual property. It has become a challenging task due to the increase in usage of the internet and digital transactions. In this scenario, DDoS attack is one of the major attacks that makes clients’ privacy questionable. It requires effective and robust attack detection and prevention techniques. Machine Learning (ML) is the most effective approach for employing cyber attack detection systems. It paves the way for a new era where human and scientific communities will benefit. This paper presents a hierarchical ML-based hyperparameter-optimization approach for classifying intrusions in a network. CICIDS 2017 standard dataset was considered for this work. Initially, data was preprocessed with the min-max scaling and SMOTE methods. The LASSO approach was used for feature selection, given as input to the hierarchical ML algorithms: XGboost, LGBM, CatBoost, Random Forest (RF), and Decision Tree (DT). All these algorithms are pretrained with hyperparameters to enhance the effectiveness of algorithms. Models performance was assessed in terms of recall, precision, accuracy, and F1-score metrics. Evaluated approaches have shown that the LGBM algorithm gives a proven performance in classifying DDoS attacks with 99.77% of classification accuracy.

Automatic identification of visual learning style in real time using raw electroencephalogram (EEG) is challenging. In this work, inspired by the powerful abilities of deep learning techniques, deep learning-based models are proposed to learn high-level feature representation for EEG visual learning identification. Existing computer-aided systems that use electroencephalograms and machine learning can reasonably assess learning styles. Despite their potential, offline processing is often necessary to eliminate artifacts and extract features, making these methods unsuitable for real-time applications. The dataset was chosen with 34 healthy subjects to measure their EEG signals during resting states (eyes open and eyes closed) and while performing learning tasks. The subjects displayed no prior knowledge of the animated educational content presented in video format. The paper presents an analysis of EEG signals measured during a resting state with closed eyes using three deep learning techniques: Long-term, short-term memory (LSTM), Long-term, short-term memory–convolutional neural network (LSTM-CNN), and Long-term, short-term memory–Fully convolutional neural network (LSTM-FCNN). The chosen techniques were based on their suitability for real-time applications with varying data lengths and the need for less computational time. The optimization of hypertuning parameters has enabled the identification of visual learners through the implementation of three techniques. LSTM-CNN technique has the highest average accuracy of 94%, a sensitivity of 80%, a specificity of 92%, and an F1 score of 94% when identifying the visual learning style of the student out of all three techniques. This research has shown that the most effective method is the deep learning-based LSTM-CNN technique, which accurately identifies a student’s visual learning style.

Deep learning models learn powerful representational spaces required for handling complex tasks. Recently, data augmentation techniques, region-level, and image-level augmentation have proved effective in significantly improving deep learning models’ generalization performance. Nevertheless, such methods may lose some critical features or are still computationally heavy (or inefficient) due to additional computation burdens. To address this issue, in this paper, we present a novel unified data augmentation method for deep learning models, namely, RandMixAugment, which effectively combines the intrinsic properties of region-level augmentation and image-level augmentation. Specifically, the proposed RandMixAugment employs automated augmentation with masking and mixing operations. Experiments are conducted on well-known CIFAR datasets (CIFAR-10 and CIFAR-100) to verify the effectiveness of the proposed scheme compared to state-of-the-art augmentation techniques. The experimental results demonstrate that the proposed RandMixAugment yields superior performance over state-of-the-art techniques on image classification tasks and further improves the performance of the baseline deep learning model by 1.2% and 2.4% on CIFAR-10 and CIFAR-100 datasets, respectively.

This paper presents an alternative methodology for the study of flyrock hazards in mining, utilizing Artificial Intelligence (AI) through machine learning by classification. By using distance as a delineator to denote the consequences of a blast, the models generated two classes of blasts: safe and unsafe. In this study, statistical learning models best suited for classification, that is, K Nearest Neighbors (KNN), Support Vector Machine (SVM), Decision Tree (DT), and Artificial Neural Networks (ANNs), were used, and their classification abilities were assessed. Machine performance was evaluated using a Confusion Matrix (sensitivity and specificity) and Receiver Operating Characteristic (ROC) curve. A higher weight was assigned to the minority class (unsafe blasts). Overfitting assessment was also performed. The Wide Neural Network (WNN) demonstrated the highest classification superiority. During training and validation, 75% sensitivity, 100% specificity, and an ROC of 0.9853 were achieved. In the test phase, perfect stratification (100 %) was maintained, with an ROC of 1. The Cubic SVM exhibited 50% sensitivity, 100% specificity, and an ROC of 0.9412 during training and validation. In the test set, it achieved 100% sensitivity, 100% specificity, and a ROC of 1. Fine KNN showed 50% sensitivity, 94.1% specificity, and an ROC of 0.7206 in the validation set. The test set displayed 100% sensitivity, 100% specificity, and an ROC of 1. Conversely, Coarse DT had a higher misclassification rate, resulting in a 25% sensitivity, 76.5% specificity, and an ROC of 0.5221 during the validation phase. In the test set, it showed 50% sensitivity, 100% specificity, and an ROC of 0.75. A feedforward neural network (FNN) was designed, trained, and demonstrated to be a highly flexible classification tool. The FNN achieved an excellent classification score of 100%. These findings demonstrate the potential for the broad applicability of machine learning through classification in addressing flyrock challenges in open-pit mines.

Effective steel surface defect classification with low computational cost is essential for online quality inspection. The challenge of this task is large intra-class differences and unclear inter-class distances shown in various surface defects. This paper proposes a novel feature extraction method based on Legendre multi-wavelet transform and Auto-Encoder network (LWT-AE) for effectively recognizing the steel surface defects categories. More precisely, the finite element approximation theory is implemented to address the strong capacity of LW bases with various regularities and vanish moments for thoroughly matching the complex geometry characteristics of the steel surface defects, which provides strict theory foundation for the feature extraction in LWT frequency domain without losing any defect information due to its orthogonality. Then, the statistical and texture parameters are utilized to sparsely extract the defect features from LWT frequency domain, resulting in removing the redundancy components corresponding to the defects. AE network is utilized to further reduce the dimension of the extracted features and automatically select the most valuable defect features. Furthermore, two classifiers (SVM and BPNN) are used to rectify the generalization ability of the proposed method. Finally, extensive experiments are conducted on two datasets to verify the effectiveness and robustness of the proposed method. The highest classification accuracies 99.44% and 95.37% compared with other methods are attained on the NEU-CLS dataset and X-SDD dataset, respectively. To summary, the proposed method not only has simple structure, but also can reliably identify different types of the steel surface defects, which provides a suitable online application technique for the actual steel surface defect classification.

In this paper, we construct a Bessel-class kernels for Support Vector Machine. This new class of kernels are proved that they are continuous and satisfy Mercer’s condition. The presented Bessel-class kernels can degenerate to Gaussian kernel in an infinite smooth case. Compared to the other kernels, these present Bessel kernels can be flexibly applied to classification and regression with fewer constants to be adjusted. Additionally, four simulation experiments including classifications and regressions, have been carried out to show the good performance of these Bessel-class kernels.

Robust methods are needed to detect how people are moving in smart public transportation systems. This paper proposes and characterizes effective means to accurately detect passengers. We analyze a public WiFi-based activity recognition (WiAR) dataset to extract human activity features from Channel State Information (CSI) data. To do so, CSI power changes caused by nearby human activity are analyzed. Our method first extracts multi-dimensional features using a Short-Time Fourier Transform (STFT) of CSI data to capture the relevant signal features. Since the environment of a transportation system changes dynamically and non-deterministically, we propose analyzing these changes with a heuristic algorithm that leverages a decision tree to automate a decision-making solution for feature selection. Principal Component Analysis (PCA) is performed before the decision tree algorithm. Reported results are compared with those obtained from the existing methods. Based on these results, we explore the effectiveness of various features such as the chirp rate, delta band power, spectral flux, and frequency of movement. This allows identifying and recommending the most effective features for the explored detection task according to observed variability, information gain, and correlation between features. The reported classification results show that using only the chirp rate estimated from CSI information as a feature, we achieve precision = 83%, True Positive $(TP)=94\%$ , True Negative $(TN)= 91\%$ and F1-score = 87%. Considering delta band power as an additional feature adds more information and allows getting higher performance with precision = 100%, $TP=97\%$ , $TN = 95\%$ and F1-score = 95%.