With the diversification trend of stock market data, constructing a stock price prediction model that can integrate multi-source data, handle high-dimensional features, and fit the complex relationships between predictive and response variables is crucial for formulating effective quantitative trading strategies. Based on this, this paper proposes a stock price prediction model that combines multi-kernel learning with Gaussian Process Neural Networks. The main innovations of the model are reflected in the following aspects: First, by using kernel functions with different domains, it effectively integrates feature information from various data sources; Second, by combining Gaussian Process Regression with neural networks, the model fully considers the impact of features from different data sources on the prediction results while addressing the curse of dimensionality and maintaining good fitting ability for complex relationships. In addition, the model can quantify the uncertainty of stock price prediction results and perform statistical inference analysis, providing more detailed auxiliary information for investors or decision-makers. Simulation analysis results show that the proposed method outperforms some classic models in predictive performance, showing strong competitiveness. Finally, in quantitative trading back-testing, considering the introduction of non-Euclidean sentiment features, the model has achieved significant superior performance in indicators such as cumulative returns, Sharpe ratio, and maximum drawdown.

A thorough analysis of developments in machine learning (ML) and deep learning (DL) technologies for skin cancer diagnosis is provided in this research. It investigates how ML and DL could improve the precision and effectiveness of melanoma, basal cell carcinoma, and squamous cell carcinoma detection. By looking at current studies, the study emphasizes the use of neural networks, convolutional neural networks (CNNs), support vector machines (SVM), random forests, and k-nearest neighbors (KNN) in the diagnosis of skin cancer. Key findings show that DL models, including VGG, ResNet, and Inception benefit from huge datasets and sophisticated data augmentation strategies to attain high accuracy, sensitivity, and specificity. The paper also discusses the challenges and limitations associated with these technologies, such as the requirement for extensive annotated datasets. The study concludes with a call for collaboration to overcome current challenges and enhance the practical application of ML and DL in skin cancer detection.

Communication is essential for humanity today and in the past. However, some individuals lack verbal communication due to their innate disability and physical losses from accidents. There are sign-language communication methods developed for such people to communicate. Artificial intelligence solutions are offered to remove the disadvantaged situations of people with disabilities due to communication in daily life. Nowadays, rapidly developing image processing and artificial intelligence methods are proper solutions for the problem focused on in this study. Convolution neural network techniques, which have become very popular recently, offer solutions to many problems. On the other hand, the YOLO algorithm shows very high performance in real-time object detection. In this study, we proposed a method for identifying the alphabets which each gesture delivers. This work studied hand detection on images and classification according to hand movements. The American Sign Language (ASL) standard was used as the sign language. The most recent version of YOLO, known as YOLOv5x, is used for gesture detection. Concentrating on the Static Sign-language problem, a study was conducted on the definition of hand movements. The letters “J” and “Z” are not included in the data set because movable hand signals are required. Apart from these two letters, a total number of 24 letters are classified. The proposed model achieved a training performance of 99.45% mAP@.5. Moreover, the proposed model has a performance of 97.9% mAP@.5 on the test dataset. The results demonstrate that the model's object detection performance is excellent. A statistical analysis of the training time shows that the training time has been drastically decreased, 4.5 hours with the current model as compared to the existing models in the literature.

COVID-19 detection through radiological examination is favoured since it is fast and produces more accurate results than the laboratory approach. However, when it has infected many people and put a strain on the healthcare system, the need for fast, automatic COVID-19 detection in patients has become critical. This study proposes to detect COVID-19 from chest X-ray (CXR) images with a machine learning approach. The main contributions of this paper are to compare two powerful deep learning models, i.e., convolutional neural networks (CNN) and the combination of CNN and Long Short-Term Memory (LSTM). In the combination model, CNN is recommended for feature extraction, and COVID-19 is classified using the features of LSTM. The dataset used in this study amounted to 4,095 CXR images, consisting of 1,400 images of normal conditions, 1,350 images of COVID-19, and 1,345 images of pneumonia. Both CNN and CNN-LSTM were executed in a similar experimental setup and evaluated using a confusion matrix. The experiment results provide evidence that the CNN-LTSM is better than the CNN deep learning model, with an overall accuracy of about 98.78%. Furthermore, it has a precision and recall of 99% and 98%, respectively. These findings will be valuable in the fast and accurate detection of COVID-19.

The rapid evolution of Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL) has profoundly influenced various domains, including portfolio optimization. In today’s dynamic and interconnected global economy, understanding the development of scientific publications in this field is crucial for both academics and practitioners. This paper aims to conduct a comprehensive bibliometric study of the scientific literature on portfolio optimization, focusing on the impact of AI, ML, and DL advancements. By analyzing key trends, influential publications, and emerging research areas, this study provides valuable insights into the progression of portfolio optimization research in the context of these transformative technologies, helping to map future directions and identify knowledge gaps in the field. This paper endeavors to present an exhaustive synthesis of the most recent advancements and innovations within the domain of portfolio optimization, particularly as influenced by progressive developments in AI, ML and DL from 1996 to 2024. Employing a rigorous bibliometric analysis, this study scrutinizes the structural and global paradigms governing this field. The analytical framework integrates several dimensions, including: (1) comprehensive dataset interrogation, (2) critical evaluation of source repositories, (3) contributions of seminal authors, (4) geographical and institutional affiliations, (5) document- centric analysis, and (6) exploration of keyword dynamics. A corpus of 745 bibliographic entries, meticulously curated from the Web of Science database, forms the basis of this inquiry, which utilizes advanced Scientometric network methodologies to extrapolate substantive research insights. The discourse culminates in a robust critique of the inherent strengths and methodological limitations, while delineating strategic avenues for future research, with the objective of steering ongoing scholarly discourse in the realm of portfolio optimization. The empirical outcomes of this study enhance the understanding of prevailing intellectual trajectories, thus laying a fortified foundation for future investigative pursuits in this critically evolving discipline.

The rapid evolution of Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL) has profoundly influenced various domains, including portfolio optimization. In today’s dynamic and interconnected global economy, understanding the development of scientific publications in this field is crucial for both academics and practitioners. This paper aims to conduct a comprehensive bibliometric study of the scientific literature on portfolio optimization, focusing on the impact of AI, ML, and DL advancements. By analyzing key trends, influential publications, and emerging research areas, this study provides valuable insights into the progression of portfolio optimization research in the context of these transformative technologies, helping to map future directions and identify knowledge gaps in the field. This paper endeavors to present an exhaustive synthesis of the most recent advancements and innovations within the domain of portfolio optimization, particularly as influenced by progressive developments in AI, ML and DL from 1996 to 2024. Employing a rigorous bibliometric analysis, this study scrutinizes the structural and global paradigms governing this field. The analytical framework integrates several dimensions, including: (1) comprehensive dataset interrogation, (2) critical evaluation of source repositories, (3) contributions of seminal authors, (4) geographical and institutional affiliations, (5) document-centric analysis, and (6) exploration of keyword dynamics. A corpus of 745 bibliographic entries, meticulously curated from the Web of Science database, forms the basis of this inquiry, which utilizes advanced Scientometric network methodologies to extrapolate substantive research insights. The discourse culminates in a robust critique of the inherent strengths and methodological limitations, while delineating strategic avenues for future research, with the objective of steering ongoing scholarly discourse in the realm of portfolio optimization. The empirical outcomes of this study enhance the understanding of prevailing intellectual trajectories, thus laying a fortified foundation for future investigative pursuits in this critically evolving discipline.

The rapid advancement of artificial intelligence (AI) has revolutionized project portfolio management (PPM), as it has in many other areas, by introducing data-driven methods that improve decision-making, risk assessment, and strategic alignment. Unlike traditional project management, which emphasizes individual project execution, PPM requires balancing multiple initiatives to optimize value creation and resource allocation. This paper presents a systematic review of scientific research on the integration of AI techniques into PPM, focusing on their applications, benefits, and challenges. The review synthesizes findings from 73 peer-reviewed studies covering a wide range of AI methodologies, such as machine learning, deep learning, neural networks, reinforcement learning, natural language processing, and hybrid optimization models. These approaches have been applied in diverse fields, including information technology, construction, healthcare, defense, energy, and telecommunications. Analysis shows that AI significantly improves project portfolio performance by predicting project outcomes, identifying interdependencies, optimizing resource allocation, and supporting adaptive strategies in dynamic environments. In addition, advanced AI tools provide project portfolio managers with predictive and prescriptive analytics, transforming PPM from reactive monitoring to proactive governance. Despite these advances, challenges remain regarding data quality, organizational readiness, and interpretability of AI-based models. Concerns about transparency, ethical implications, and integration with existing management frameworks also hinder wider adoption. However, recent developments indicate a growing trend toward hybrid systems that combine AI with traditional decision-making models, increasing both accuracy and practical applicability. This review contributes to theory and practice by synthesizing current knowledge, highlighting research gaps, and identifying emerging directions such as the use of large language models, ensemble methods, and sustainability-focused project portfolio optimization. The findings highlight the transformative potential of AI in advancing PPM and provide valuable insights for researchers and practitioners seeking to design smarter, more adaptive, and more sustainable project portfolio management strategies.

The increasing complexity and frequency of cyberattacks have made accurate and reliable intrusion detection systems (IDSs) essential for modern network security. In this study, an explainable triple-hybrid deep learning framework is proposed for binary intrusion detection using the CICIDS2017 dataset. The proposed architecture integrates three complementary branches, namely a Transformer encoder, a bidirectional long short-term memory (BiLSTM) network, and a multilayer perceptron (MLP), to capture global feature interactions, sequential dependencies, and nonlinear discriminative patterns from network traffic data. To enhance adaptive representation learning, the framework employs a branch-gating mechanism and a fusion-gating module before final classification. The model was evaluated in a Benign-versus-Attack setting using 5-fold stratified cross-validation and assessed through accuracy, precision, recall, F1-score, ROC-AUC, confusion matrices, and explainability analysis. Experimental results showed strong and stable performance across folds, with a mean validation accuracy of 97.18%, a best-fold accuracy of 97.43%, and a mean ROC-AUC of 0.9975. LIME-based explanations further improved transparency, confirming the framework as an effective and interpretable solution for binary intrusion detection.

The swift integration of intelligent technologies at higher institutions of learning has greatly improved efficiency of operations, learning conditions, and administration. But the evolution of cybersecurity issues presented by the integration of Internet of Things (IoT) devices, cloud-based systems, and interconnected systems has complicated and shifted the complexity of these issues, which old and standard methods of periodic auditing cannot effectively tackle. The present paper suggests an AI-based Cyber Risk Auditing Framework (AI-CRAF) of continuous and real-time risk assessment in smart educational campuses. The framework combines sophisticated machine learning and deep learning models to identify the threats, anomalies, and dynamic risk assessment with references to vulnerabilities to the system and their potential impact. The suggested model is tested on a big data set of 1,247,334 events within 12 months that contains various attack cases and regular operations. The experimental values indicate a high detection accuracy of 96.2 %, a true detection rate of 94.8 %, a low false positive rate of 2.1 % and an AUC-ROC value of 0.978. Also, the framework shortens 97.1 the time spent on an average incident response by 41.9 minutes to 1.2 minutes as compared to conventional methods.

Malaria remains one of the most dangerous infectious diseases in tropical and sub-Saharan regions, requiring efficient, accurate, and interpretable diagnostic systems. This paper presents a hybrid system that combines machine learning and deep learning techniques to classify malaria-infected blood cells from microscopic images. The approach integrates artificial features such as texture features from the Gray Level Co-occurrence Matrix (GLCM), morphological features, and statistical features with deep features extracted from pre-trained convolutional neural networks like ResNet50 and VGG16. After preprocessing steps, including gamma correction, HSV color space transformation, and contrast-limited adaptive histogram equalization (CLAHE), the cells were segmented using local entropy thresholding and filtered to remove noise smaller than 125 pixels. The combined handcrafted and deep features were classified using Support Vector Machine (SVM), Random Forest (RF), and XGBoost classifiers, evaluated through 10-fold cross-validation. Results demonstrate that the hybrid model significantly improves performance over methods based on single features. The XGBoost classifier achieved the highest accuracy at 95.4%, with precision of 95.1%, recall of 94.8%, and an AUC of 0.99.