Electronics & Computer Engineering هندسة الإلكترونيات والحاسوب
Permanent URI for this collection
Browse
Recent Submissions
- Itemنحو استخراج مترادفات للغة العربية ألياً(Al-Quds University, 2025-01-11) Eman Abed Al-Kareem Mousa Naser; ايمان عبد الكريم موسى نصرSynonyms extraction gains special attention as synonyms are essential in improving Natural Language Processing (NLP) application performance. The Lexical Substitution (LS) task is utilized for Synonym extraction, which generates a set of equivalent substitutions (i.e., synonyms) to the target word or phrase in a sentence that saves the sentence's meaning. This task can enhance writing, language understanding, and NLP models and address ambiguity. Recently, LS has attracted much attention in many languages. Despite the richness of Arabic vocabulary, limited research has been performed on the LS task due to the lack of annotated data. To bridge this gap, we present the first Arabic LS benchmark dataset, AraLexSubD for benchmarking LS pipelines. AraLexSubD is manually built by eight native Arabic speakers and linguists (six linguist annotators, a doctor, and an economist) who annotate the 630 sentences. AraLexSubD covers three domains: general, finance, and medical. It encompasses 2476 substitution candidates ranked according to their semantic relatedness. We also present an Arabic LS pipeline, AraLexSubPro, which offers different techniques for generating, selecting, and ranking substitutions. To make a thorough comparison, AraLexSubPro uses four different methods as baselines to generate substitute candidates for the target words: a synonym dictionary-based approach using Arabic Word Net (AWN), a pre-trained language model-based approach (AraBERT), AraBERT dropout (partial masking), and a hybrid approach between AraBERT and AWN. The results showed that the hybrid approach achieved the best results compared to the other approaches. The generated substitutions are filtered and then ranked based on six high-quality features to compare thoroughly: word similarity, word frequency, BERT prediction order (BERT probability), BERT-based language model (Loss), BERT similarity, and the BERTscore. The substitutions are then reranked based on our AraLexSubPro ranker. Additionally, an error analysis of the experiment is reported. To evaluate the AraLexSubPro pipeline, we use our first benchmark dataset for the Arabic LS task AraLexSubD dataset, which can automatically evaluate the Arabic LS systems. To our knowledge, this is the first study on Arabic lexical substitution. The results were encouraging and fundamental for Arabic LS research. To speed up research on this field, we have put the AraLexSubD data on GitHub at the following link: https://github.com/karajah2024/Arabic-Lexical-Substitution.git
- ItemDiscovering Gene Associations Across Diseases Using a Knowledge-based Machine Learning Approach(Al-Quds university, 2024-11-04) Emma Mamdouh Jeries Qumsiyeh; ايما ممدوح جريس قمصيةComplex diseases such as diabetes, Alzheimer's, and cancer are influenced by a combination of genetic, lifestyle, and environmental factors that do not follow straightforward inheritance patterns. Biological systems are immensely complex and heterogeneous. To resolve the enigmas surrounding these systems, extensive research provides huge amounts of biological data. In this thesis and in our first study, a novel approach called GediNET was developed to integrate prior biological knowledge into disease-associated gene groups. GediNET employs a Grouping, Scoring, and Modeling (G-S-M) approach to identify top-performing gene groups, which are then used to train a machine-learning model. Following the data exploration and preprocessing steps, various classification models were built with 100-fold Monte Carlo Cross-Validation, and the performance of these models was evaluated. By applying Disease-Disease Association (DDA) based machine learning, GediNET uncovered new relationships between diseases, improving diagnosis, prognosis, and treatment approaches. In the second study, GediNETPro, an advanced version of GediNET, was developed. This version utilizes Cross-Validation (CV) information and clustering techniques, such as K-means, to identify patterns of disease group associations. GediNETPro provides visualization tools, like heatmaps and in-depth analysis of disease group clusters, offering insights for developing effective diagnostic interventions. The third study leveraged molecular-level data to develop effective methods for predicting Disease-Disease Associations (DDAs). A statistical technique was developed by employing the G-S-M-P model of GediNETPro to compute semantic similarity metrics between diseases. The semantic approach detects representative diseases within clusters and establishes a semantic relationship between the disease under investigation and other diseases. The studies presented in this thesis contribute to understanding disease complexity, uncovering disease associations, and identifying potential biomarkers and drug targets
- Itemنحو استخراج مترادفات للغة العربية ألياً(Al-Quds University, 2025-01-11) ايمان عبد الكريم موسى نصر; Eman Abed Al-Kareem Mousa Naserث، وضعنا البيانات AraLexSubD على منصة GitHubعلى الرابط التالي: https://github.com/karajah2024/Arabic-Lexical-Substitution.git Synonyms extraction gains special attention as synonyms are essential in improving Natural Language Processing (NLP) application performance. The Lexical Substitution (LS) task is utilized for Synonym extraction, which generates a set of equivalent substitutions (i.e., synonyms) to the target word or phrase in a sentence that saves the sentence's meaning. This task can enhance writing, language understanding, and NLP models and address ambiguity. Recently, LS has attracted much attention in many languages. Despite the richness of Arabic vocabulary, limited research has been performed on the LS task due to the lack of annotated data. To bridge this gap, we present the first Arabic LS benchmark dataset, AraLexSubD for benchmarking LS pipelines. AraLexSubD is manually built by eight native Arabic speakers and linguists (six linguist annotators, a doctor, and an economist) who annotate the 630 sentences. AraLexSubD covers three domains: general, finance, and medical. It encompasses 2476 substitution candidates ranked according to their semantic relatedness. We also present an Arabic LS pipeline, AraLexSubPro, which offers different techniques for generating, selecting, and ranking substitutions. To make a thorough comparison, AraLexSubPro uses four different methods as baselines to generate substitute candidates for the target words: a synonym dictionary-based approach using Arabic Word Net (AWN), a pre-trained language model-based approach (AraBERT), AraBERT dropout (partial masking), and a hybrid approach between AraBERT and AWN. The results showed that the hybrid approach achieved the best results compared to the other approaches. The generated substitutions are filtered and then ranked based on six high-quality features to compare thoroughly: word similarity, word frequency, BERT prediction order (BERT probability), BERT-based language model (Loss), BERT similarity, and the BERTscore. The substitutions are then reranked based on our AraLexSubPro ranker. Additionally, an error analysis of the experiment is reported. To evaluate the AraLexSubPro pipeline, we use our first benchmark dataset for the Arabic LS task AraLexSubD dataset, which can automatically evaluate the Arabic LS systems. To our knowledge, this is the first study on Arabic lexical substitution. The results were encouraging and fundamental for Arabic LS research. To speed up research on this field, we have put the AraLexSubD data on GitHub at the following link: https://github.com/karajah2024/Arabic-Lexical-Substitution.git
- ItemHybrid Anomaly Based Android Malware Detection Using Deep Neural Networks(Al-Quds University, 2023-08-29) Maher George Mousa Maria; ماهر جورج موسى مارياA malware detection system for mobile devices contributes to the field of computer security. Cybersecurity is a major current problem mainly motivated by the growing number of malwares; data loss due to computer breaches cost a great loss. In addition ethical problems. Due to the popularity of smartphones and tablets, mobile devices are becoming the target of malware and cyberattacks. It is therefore essential to explore new ways to prevent, detect and counter cyberattacks. In these detection mechanisms, machine learning is used to create classifiers that determine whether an application is compromised. The advantage of a neural network is that it allows you to adapt to new situations. Therefore, we used this new technology to be able to identify types of malicious behavior and to be able to generalize it to future malicious programs. The goal of this thesis is to propose a malware detection model on Android based on deep neural networks classification driven by sets of hybrid features. We reviewed and classified existing methods into two groups: the static methods which consist of examining the code of the mobile application and the dynamic methods which analyze the behavior of an application when it is running on a mobile terminal. Our goal is to use these two methods to take advantage of the both groups. To do this, we used the hybrid database “AMD” composed of 85 features. We are also conducted an experiment plan composed of hundreds of trainings in order to adjust the values of the hyperparameters improving the learning on this dataset as well as to select the most relevant remaining features, through this thesis, we work according to the most effective features from the AMD Dataset. And to improve detection accuracy that have time-dependent frequencies such as attacks, three new input features (s_sessiontime, r_sessiontime, and sr_sessionime) are devised by aggregating the flows based on source, destination, and timestamp attributes using a time window of one minute. Also, after preprocessing the input features, the most important 45 input features are selected. Moreover, the model’s parameters are learned using many multiclass labeled flows from the AMD dataset. The hyperparameters of the model are optimized for best performance in terms of accuracy, recall, precision, and training time of the model. The experimental results confirmed the high performance of the proposed model when tested from the “AMD” dataset. In addition, the optimal model architecture consists of one input layer, three hidden layers and one output layer. The model achieved an accuracy of 99.8 %, a false positive rate of less than 1%, and an area under the receiver operating characteristic curve (ROC-AUC) of 0.999. Also, the detection accuracy of the multiclass classifier is 99.6% When the proposed model is compared with other recent models in literature, that was evaluated on similar datasets like” AMD”, the experimental results show that the proposed model outperforms other models in terms of precision and recall.
- ItemIntelligent Vertical Handover Model based on Neuro-Fuzzy for LTE and WiFi networks (IVH-NF LTE/WiFi)(Al-Quds University, 2022-05-15) Niveen Omar Saleh Jaffal; نفين عمر صالح جفالساهم النمو السريع للإنترنت وتطور معدات المستخدم (UE) مع وجود الواجهات المتعددة للشبكات، في توفير المزيد من جودة الخدمات وتمكين المستخدمين من اختيار الاتصال بشبكات الوصول المتعددة، ذات الخصائص المختلفة، والانتقال من شبكة إلى أخرى. تختلف الشبكات ذات التطور طويل الأمد (LTE) من عائلة الشبكات الخلوية والشبكة المحلية اللاسلكية (WiFi) من عائلة IEEE802.x في منطقة التغطية, ووقت الاستجابة، ومعدل البيانات، وقوة الإشارة، والتكلفة، وعرض النطاق الترددي، وكل واحدة من هذه الشبكات مصممة لدعم مجموعة مختلفة من UE وخدمات محددة، وبالتالي لِتغّلب على هذه الاختلافات في الشبكات اللاسلكية التكميلية من خلال دمج هذه التقنيات المختلفة في NGWN [1]. لجعل خدمة الاتصال متوفرة دائما ل UE بأفضل معايير الجودة بين هاتين الشبكتين غير المتجانستين، فإن عملية التسليم الرأسي ضرورية لتحقيق ذلك، وهو التبديل بين نقاط التعلق أو المحطات الأساسية ضمن تقنيات الشبكة. مع تطور التقنيات في الوقت الحاضر، لا يزال البحث عن خوارزميات التسليم العمودي في شبكات سيناريوهات التنقل المختلفة اليوم يمثل مجالًا صعبًا للبحث في المستقبل. في هذه الدراسة، اقترحنا خوارزمية ذكية تسمى IVH_NF متوافقة مع UE لقرار التسليم العمودي مع أربعة متغيرات إدخال (قوة الإشارة المستلمة RSS، معدل استهلاك الطاقة PC، سرعة التنقل MSو المسافة D) التي تجمع من الشبكة المحيطة لـ UE بين شبكتين غير متجانستين:WiFi و LTE باستخدام تقنية ANFIS التي تعتمد على Takagi – Sugeno FIS التي تجمع بين كل من الشبكات العصبية ومبادئ المنطق الضبابي، فهي تستفيد من كليهما في إطار واحد، ويبدأ مبدأ العمل من السماح للشبكة العصبية بالتعلم من أجل ضبط متغيرات نظام الاستدلال الضبابي (FIS) باستخدام مجموعة البيانات التدريب Data Set لإنشاء قواعد عضوية غامضة Membership Function Rules تكون مسؤولة عن اختيار الشبكة المثلى مع تحقيق أفضل جودة خدمة من حيث الإنتاجية المحسنة, وتقليل عدد عمليات التسليم غير الضرورية والتأخير النهائي. تتضمن نتائج المحاكاة التحقق أولاً من دقة التصميم الرياضي IVH_NF باستخدام محاكي MATLAB من خلال أدوات قياس الخطأ القياسية مع معايير إحصائية مثل RMSE و MAE و R2 حيث كشفت نتائج المحاكاة أن التصميم الرياضي يحقق مستوى عالٍ من الدقة، بعد ذلك تم العديد من تنفيذ سيناريوهات في عملية التسليم الرأسية من LTE إلى شبكات WiFi وتم إجراء تجارب محاكاة مكثفة باستخدام محاكي Apache NetBeans الذي يقرأ قواعد القرار لتسليم العمودي كملف txt من MATAB لدراسة تأثير سرعة UE ومعدل البتBit Rate للتطبيق على مقياس الأداء الذي يؤثر على جودة الخدمة، مثل الإنتاجية و زمن التأخير EED وعدد عمليات التسليم لسيناريوهات التسليم العمودي. تؤكد النتائج التي تم الحصول عليها من السيناريوهات أن القيم مقبولة ومع مقارنة النموذج المقترح مع بعض الخوارزميات ذات الصلة، نستنتج أن النموذج المقترح يحقق نتائج أفضل من حيث تحسين الإنتاجية, تقليل EED بنسبة 58٪ من RSS[2] وخفضت عدد عمليات التسليم بنسبة 33.3٪ من ANFIS[3] و 60% من .FL [4]