Egyptian Informatics Journal最新文献

{"title":"Graph-based temporal anomaly detection with self-supervised contrastive learning and dynamic adaptive thresholding for acoustic howling suppression","authors":"Xiaoqian Fan ,&nbsp;Francisco Hernando-Gallego ,&nbsp;Diego Martín ,&nbsp;Mohammad Khishe","doi":"10.1016/j.eij.2026.100892","DOIUrl":"10.1016/j.eij.2026.100892","url":null,"abstract":"<div><div>Acoustic howling due to feedback loops in audio systems is a major challenge in such fields as hearing aids or public address systems. Traditional approaches such as notch filters and adaptive feedback cancellation often have limitations such as lack of adaptability in dynamic environments, and a need for a large amount of labelled data. To overcome these shortcomings, a new deep learning approach, Dynamic Adaptive Thresholding and Self-Supervised Contrastive Learning for Graph-based Temporal Anomaly Recognition (GTAD-CL), is proposed in this paper. By representing audio signals as graphs, GTAD-CL uses graph neural networks to represent complex spatial–temporal patterns to detect howling with high precision as an anomaly. Self-supervised contrastive learning removes the requirement of having labeled datasets which improves the scalability and generalization of the AI models. A dynamic adaptive thresholding mechanism guarantees robust performance under different acoustic conditions, e.g. low signal to noise ratio environments. Integrated with neural filtering in real time, GTAD-CL makes howling suppression easy. Experimental results on a 100-hour custom dataset and six public benchmarks indicate that GTAD-CL has a precision of 0.92 (compared to 0.88, the best baseline, HybridAHS, showing a gain of 4.5%), recall of 0.90 (compared to 0.85, a gain of 5%) and F1-score of 0.91 (compared to 0.865, a gain of 4.5%). In suppression quality GTAD-CL achieves a PESQ score of 3.02 (compared to 2.68 for HybridAHS, i.e. ∼12.7% better), and a STOI of 0.90 (compared to 0.86, i.e. ∼4.7% better). Moreover, GTAD-Cl runs with a real-time factor of 0.36× which is better than HybridAHS’s 0.42× (approx. 14% faster). These results give validation to GTAD-CL as a powerful, scalable, and low-latency solution and high-fidelity solution that is superior to state-of-the-art results for varying acoustic scenarios.</div></div>","PeriodicalId":56010,"journal":{"name":"Egyptian Informatics Journal","volume":"33 ","pages":"Article 100892"},"PeriodicalIF":4.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146078372","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"KDLog: a selective knowledge distillation approach for sequential log anomaly detection","authors":"Hailong Cheng ,&nbsp;Shi Ying ,&nbsp;Xiaoyu Duan ,&nbsp;Wanli Yuan","doi":"10.1016/j.eij.2025.100879","DOIUrl":"10.1016/j.eij.2025.100879","url":null,"abstract":"<div><div>Log anomaly detection is a critical task for ensuring the reliability of complex systems. However, existing methods often suffer from poor adaptability and substantial retraining overhead as log data evolve. This paper introduces a novel framework called KDLog, a knowledge-distillation-based approach that enables accurate and efficient log anomaly detection in dynamic environments. KDLog employs a two-stage selective-distillation mechanism, in which a lightweight student model is trained using the high-confidence outputs generated by a teacher model, effectively preventing negative knowledge transfer. Compared with state-of-the-art methods, KDLog improves overall accuracy by 4.5%, F1-score by 4.3%, and recall by 3.3% on average across real-world datasets (HDFS and BGL). Moreover, it reduces model update time by 60–78% and achieves a smaller model size, by up to 50%, compared with deep learning baselines such as DeepLog and LogAnomaly. Statistical significance tests confirm the robustness of these improvements. Unlike prior methods, KDLog also demonstrates strong resilience to unseen log patterns, with less than a 4% performance drop under simulated log-template drift. These gains make KDLog a scalable and practical solution for real-time anomaly detection, effectively bridging the gap between high-performance learning and operational efficiency in production environments.</div></div>","PeriodicalId":56010,"journal":{"name":"Egyptian Informatics Journal","volume":"33 ","pages":"Article 100879"},"PeriodicalIF":4.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884767","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Adaptive acoustic feedback control in aphasia Therapy: A Graph-Based learning approach for Unintended resonance suppression in Mandarin (Chinese)-Speaking aphasic patients","authors":"Jiayi Zhang ,&nbsp;Nuria Serrano ,&nbsp;Francisco Hernando-Gallego ,&nbsp;Mohammad Khishe","doi":"10.1016/j.eij.2026.100908","DOIUrl":"10.1016/j.eij.2026.100908","url":null,"abstract":"<div><div>Aphasia therapy for Mandarin-speaking patients presents distinct challenges due to the language’s tonal characteristics and the presence of unforeseen vocal resonance, which reduces intelligibility and distorts tone contours. Current automatic speech feedback systems face challenges managing such distortions, especially in real-time and customized clinical contexts. This paper develops a novel framework, named graph-based adaptive acoustic feedback control (GA-AFC), that integrates graph neural networks (GNNs) with reinforcement learning (RL) to model and suppress articulation-resonance mismatches in aphasic speech in a dynamic manner. Unlike black-box automatic speech recognition (ASR) and traditional autoregressive models, GA-AFC constructs an articulation-resonance graph based on acoustic features such as harmonicity, pitch, energy, and Mel-frequency cepstral coefficients (MFCCs). The system utilizes GNN encoders to capture phoneme-tonal transitions and employs an RL policy to adapt acoustic feedback in real-time. Experimental evaluations on three benchmark Mandarin datasets, i.e., Common Voice (Mandarin), AISHELL-1, and HKUST, demonstrate that GA-AFC achieves substantial improvements in both fluency enhancement and recognition accuracy. In the context of aphasic speech, the model achieves an average word error reduction (WER) of 17.2% relative to Wav2Vec 2.0 and 30.1% relative to DeepSpeech, alongside a 14.8% improvement in tone classification accuracy on the HKUST corpus. Regarding resonance suppression, GA-AFC logs a spectral deviation of baseline systems by 28.6%, achieving a MOS score of 4.4 (±0.3) in subjective listening tests, which surpasses all comparative models. Moreover, the system demonstrates rapid convergence, with adaptation times of less than 20 s and feedback latencies of under 140 <em>ms</em>, making it suitable for real-time clinical use. The findings indicate that GA-AFC provides a responsive, adaptable, and clinically applicable framework for customizable speech feedback in Mandarin aphasia therapy, proposing a novel approach to tone- and resonance-sensitive neural interventions in speech rehabilitation.</div></div>","PeriodicalId":56010,"journal":{"name":"Egyptian Informatics Journal","volume":"33 ","pages":"Article 100908"},"PeriodicalIF":4.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147396705","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A novel framework for phishing detection based on backward recursive feature selection","authors":"Esraa Abu Elsoud ,&nbsp;Malek Bahroush ,&nbsp;Esraa Al Henawi ,&nbsp;Asma Alnajjar ,&nbsp;Abdelarahman Makhoul ,&nbsp;Omar Owedat ,&nbsp;Ismaiel Barqan ,&nbsp;Adel Alsaqqar","doi":"10.1016/j.eij.2026.100922","DOIUrl":"10.1016/j.eij.2026.100922","url":null,"abstract":"<div><div>Feature selection plays a crucial role in enhancing the accuracy and efficiency of phishing detection systems. Various innovative approaches have been proposed to optimize feature selection, which is essential for improving model performance and reducing false positives. This study presents a structured method that combines email parsing with hashing-based feature extraction to enhance the accuracy of phishing detection models. During the parsing phase, it identifies key elements like linguistic patterns, metadata, embedded URLs, and attachments, ensuring that only the most relevant information is used for further analysis. Next, a hashing technique is employed to convert high-dimensional textual data into fixed-size feature vectors, which helps maintain important meanings while simplifying the data. On the other hand, this research presents a hybrid method for feature selection that merges Recursive Feature Elimination (RFE) with Backward Elimination (BE) to improve the efficiency of ML models. The proposed framework, Recursive Backward Elimination (RBE), enhances detection precision, decreases dimensionality, and lowers computational expenses. Results indicate significant improvements across classifiers, with the Ensemble technique reaching the peak accuracy of 90.7%. The RBE is designed to power strong machine learning pipelines that can accurately detect phishing attempts, all while ensuring data privacy and scalability.</div></div>","PeriodicalId":56010,"journal":{"name":"Egyptian Informatics Journal","volume":"33 ","pages":"Article 100922"},"PeriodicalIF":4.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147396701","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Probabilistic History-based distributed sensing protocol in cognitive radio networks","authors":"Ramzi Saifan ,&nbsp;Rami Al-zyadat ,&nbsp;Mohammed Hawa ,&nbsp;Iyad Jafar ,&nbsp;Samah Rahamneh","doi":"10.1016/j.eij.2026.100885","DOIUrl":"10.1016/j.eij.2026.100885","url":null,"abstract":"<div><div>Cognitive radio technology allows opportunistic access to underutilized licensed radio spectrum by dynamically sensing and accessing this scarce resource. Existing cognitive radio network (CRN) protocols may suffer from limited number of transmission channels, conflicts among secondary users (SUs), and a lack of awareness about potentially superior available channels. This paper introduces a decentralized protocol to enhance CRN performance, allowing seamless acquisition of spectrum bands. Unlike traditional protocols with information exchange overhead, our proposed method does not require communication between SUs, ensuring high performance and minimal interference.</div><div>The proposed protocol is called Probabilistic History-based Distributed Sensing Protocol in Cognitive Radio Networks (PHDS-CRN), and it is designed to address the limitations of existing protocols. This protocol offers a fully distributed approach to spectrum sensing and channel allocation, enabling SUs to efficiently utilize available spectrum bands while minimizing interference with primary users (PUs) and other SUs. By categorizing spectrum bands into distinct groups and employing a three-phase decision process, PHDS-CRN optimizes channel access in CRNs. Our experimental evaluation demonstrates the superior performance of PHDS-CRN compared to existing methodologies. Under 100% load conditions, our proposed method achieves high channel access rate, while significantly reducing settling time and interference time.</div></div>","PeriodicalId":56010,"journal":{"name":"Egyptian Informatics Journal","volume":"33 ","pages":"Article 100885"},"PeriodicalIF":4.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145977568","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Reconstruction of project quality assessment through a data-driven machine learning model","authors":"Ching-Lung Fan","doi":"10.1016/j.eij.2026.100900","DOIUrl":"10.1016/j.eij.2026.100900","url":null,"abstract":"<div><div>Public construction quality in Taiwan is commonly assessed through committee-based inspections, yet the resulting scores are often subjective and heavily concentrated within narrow grading ranges. To address this limitation, this study proposes a data-driven framework that integrates Principal Component Analysis (PCA) with a Multilayer Neural Network (MNN) to reconstruct objective and discriminative ranges of project quality scores. Using 962 inspection records from the Public Construction Intelligence Cloud (PCIC), PCA is first applied to reduce 499 defect items into 13 representative serious defects, mitigating multicollinearity and retaining the most informative quality indicators. These defects, together with the project contract amount and construction progress, are then used as inputs to an optimized MNN classifier. A systematic hyperparameter search and stratified 10-fold cross-validation are employed to ensure robust model generalization. Based on the learned relationships, new grading thresholds are derived: A+ (86–100), A (83–86), A– (80–83), and B+ (&lt;80). The proposed PCA–MNN framework achieves an overall accuracy of 95% and significantly alleviates the extreme class imbalance observed in the original scoring scheme. Results demonstrate that the reconstructed ranges provide a more balanced, interpretable, and objective representation of project quality, enabling fairer multi-class evaluation and supporting more reliable decision-making in public construction quality management.</div></div>","PeriodicalId":56010,"journal":{"name":"Egyptian Informatics Journal","volume":"33 ","pages":"Article 100900"},"PeriodicalIF":4.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146078374","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An accurate similarity-based model for movie rating prediction and recommendation using an uncertainty score","authors":"Youssef Hanyf ,&nbsp;Hassan Silkan ,&nbsp;Abdellatif Dahmouni ,&nbsp;Abdelkaher Ait Abdelouahad","doi":"10.1016/j.eij.2025.100860","DOIUrl":"10.1016/j.eij.2025.100860","url":null,"abstract":"<div><div>This paper presents a novel method for movie rating prediction and recommendation systems based on similarity between movies with an uncertainty score to control prediction confidence. The two traditional recommendation approaches, namely collaborative filtering and content-based, rely on the concept of similarity between movies and users. Although similarity plays a crucial role in recommendation systems, it has not been sufficiently explored in existing research. To bridge this gap, we propose a dissimilarity function for movies based on a thorough analysis of movie features. We also introduce an uncertainty score that quantifies the confidence in predictions based on the dissimilarity between the unseen movie and the nearest rated movie. The proposed method uses the uncertainty score for two purposes. First, it adjusts the predicted rating by shifting it toward the user’s mean rating when the uncertainty exceeds a predefined threshold. Second, it prioritizes recommendations based on the uncertainty score, allowing the system to recommend only movies with high prediction certainty. The experimental results show that the proposed method is significantly accurate at lower uncertainty thresholds (≤12%). Furthermore, the method also performs well in top-K movie recommendations, providing consistent performance regardless of the number of recommended movies when uncertainty is low. The proposed method is also compared with state-of-the-art machine learning models, such as Support Vector Machine Regression, Random Forest Regressor, and Gradient Boosting Regressor. The comparison shows that our approach outperforms these models at low uncertainty levels and provides more reliable and accurate recommendations.</div></div>","PeriodicalId":56010,"journal":{"name":"Egyptian Informatics Journal","volume":"33 ","pages":"Article 100860"},"PeriodicalIF":4.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145791744","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Fostering Creative sports talents with transformer models for inclusive financial","authors":"Yang Gao ,&nbsp;Wenjie wang ,&nbsp;Yangyang Li","doi":"10.1016/j.eij.2025.100838","DOIUrl":"10.1016/j.eij.2025.100838","url":null,"abstract":"<div><div>The study introduces a novel sports analytics approach and, for the first time, applies the TabTransformer model to predict attendance at fitness classes. The main objective is to uncover potential sports talent and create in-depth financial planning based on attendance patterns. Compared to the deep model and traditional model, the TabTransformer performs better with an accuracy of 0.710, precision of 0.738, recall of 0.707, F1 score of 0.722, and an AUC-ROC of 0.818. This is because the model can make use of textual embeddings to handle categorical features and linear transformations for numerical features, which are able to capture complex interactions between data. The results depict the high ability of the model to identify committed members in well-attended groups (e.g., Aqua and HIIT), but the model’s moderate recovery in poor-attendance groups (e.g., Strength and Cycling) directs us towards further investigation of barriers to access. These insights pave the way for designing targeted interventions and inclusive financial strategies, including membership subsidies and flexible schedules. Despite limitations such as the moderate size of the dataset and the lack of financial features, this research lays a strong foundation for the application of Transformer models in sports analytics. Ultimately, this study emphasizes the importance of using Transformer-based analytics to generate creative and equitable outcomes in fitness programs and is a step forward in identifying talent and promoting inclusion in sports.</div></div>","PeriodicalId":56010,"journal":{"name":"Egyptian Informatics Journal","volume":"33 ","pages":"Article 100838"},"PeriodicalIF":4.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145791747","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"IFC: Editorial","authors":"","doi":"10.1016/S1110-8665(26)00048-4","DOIUrl":"10.1016/S1110-8665(26)00048-4","url":null,"abstract":"","PeriodicalId":56010,"journal":{"name":"Egyptian Informatics Journal","volume":"33 ","pages":"Article 100931"},"PeriodicalIF":4.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147538518","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Morphological edge enhancement and bottleneck layers for robust underwater object detection","authors":"Vasanthi Ponduri ,&nbsp;Sumanth kumar Panguluri ,&nbsp;D Vidyanadha Babu ,&nbsp;Jammisetty Yedukondalu ,&nbsp;Lakshmi Prasanna Kothala ,&nbsp;Chandana Gowri Doddi","doi":"10.1016/j.eij.2026.100923","DOIUrl":"10.1016/j.eij.2026.100923","url":null,"abstract":"<div><div>Underwater object detection is particularly challenging due to the inherent distortions, scattering, and light attenuation present in aquatic environments, which degrade visual clarity and impact detection accuracy. This paper outlines a comprehensive approach to enhance object detection in such challenging conditions. The proposed methodology combines morphological edge enhancement techniques with an efficient detection model featuring bottleneck layers. The initial phase involves the acquisition of a meticulously labeled dataset comprising underwater images containing objects of interest. Prior to model training, a critical preprocessing step is undertaken to rectify underwater distortions, encompassing tasks like color correction and contrast enhancement. To further fortify the proposed method adaptability to various conditions in underwater, the dataset is enriched through augmentation, introducing variations in lighting conditions, water clarity, and object poses. Bottleneck layers act as information bottlenecks, reducing the dimensionality of features while simultaneously enhancing their depth. This transformation not only compresses information but also mitigates computational overhead, thereby facilitating efficient object detection. This proposed model undergoes experimental validation on the underwater dataset, achieving significantly higher metrics such as a mean average precision (mAP) of 85.1%, precision of 84.4%, and recall of 79.9%. These experimental findings strongly indicate that the suggested method surpasses current models in its ability to detect exceedingly underwater objects effectively.</div></div>","PeriodicalId":56010,"journal":{"name":"Egyptian Informatics Journal","volume":"33 ","pages":"Article 100923"},"PeriodicalIF":4.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147396702","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}