http://dspace.univ-mascara.dz:8080/jspui/handle/123456789/15 2026-08-26T02:00:37Z 2026-08-26T02:00:37Z SALMA, Hanane http://dspace.univ-mascara.dz:8080/jspui/handle/123456789/1484 2026-07-16T10:57:12Z 2026-07-16T00:00:00Z

Titre: Optimization of Dynamic Analytical Queries Using Advanced Methods Auteur(s): SALMA, Hanane Résumé: In modern data management systems, redundant computations and data duplication are the current challenges. This occurs because of analytical queries which often share common computational components called common subexpressions, significantly increasing processing time and costs. This phenomenon is present in various data processing paradigms, such as transactional (OLTP), analytical (OLAP), and real time processing (RTAP) systems.Recent studies demonstrated that over 18% of queries in cloud environments, such as Alibaba, contain computational duplication. Two queries are considered duplicates if they share the same subexpressions or have structural similarities. The challenge of discovering and exploiting these subexpressions is known as the common subexpression selection problem, a complex NP-complete problem.This problem is relevant to several key areas of database performance optimization, including multiquery optimization, materialized views, query rewriting, and big data and cloud computing.Physical scene selection is a key solution in this field, aiming to choose the best set of subexpressions that achieve performance optimization while considering storage, update, and reuse constraints. Although this area has been studied since the 1990s through more than 180 research papers, most traditional solutions remain limited to static environments and small query sizes.Today, however, modern applications require dynamic and large-scale processing. Systems like BIGSUBS[81] have proved their ability to detect more then thousands of subexpressions, reducing execution time and cost. Therefore, developing smarter and more flexible solutions has become essential. This thesis aims to propose solutions based on deep learning, using efficient data structures such as graphs, with costbased models to estimate the benefit of common expression selection. The goal is to build scalable optimization strategies compatible with modern and complex query processing systems.

2026-07-16T00:00:00Z DJELLAL SERANDI, Mohamed http://dspace.univ-mascara.dz:8080/jspui/handle/123456789/1481 2026-07-15T13:44:17Z 2026-07-15T00:00:00Z

Titre: Towards a Classification and Detection Model for Medical Diagnosis Based on Machine Learning Algorithms Auteur(s): DJELLAL SERANDI, Mohamed Résumé: High-throughput DNA microarray technology produces high-dimensional gene ex- pression data characterized by a large number of irrelevant and redundant genes, which exacerbates the curse of dimensionality and negatively impacts disease clas- sification performance. Consequently, effective feature selection is a critical step in gene expression analysis. This thesis investigates bio-inspired and deep learning–based feature selection strategies for gene expression data to improve classification performance through comparative analyses between single-objective and multi-objective Differential Evo- lution frameworks, Differential Evolution and autoencoder-based feature learning approaches, and Differential Evolution and the Slime Mould Algorithm, highlighting the benefits of multi-objective optimization in balancing the mean squared residual (MSR) and feature subset compactness; moreover, a novel clustering-based multi- objective Differential Evolution framework (CBDE) is proposed to enhance search efficiency and identify compact and biologically meaningful gene subsets. The proposed methods are evaluated on multiple benchmark microarray datasets using state-of-the-art classifiers, including Support Vector Machine (SVM), Naive Bayes (NB), K-Nearest Neighbors (KNN), Decision Tree (DT), Random Forest (RF), and Logistic Regression (LR). Experimental results demonstrate that the applica- tion of feature selection significantly improves convergence behavior and classification performance, while promoting a diverse set of optimal solutions. Moreover, the proposed approaches consistently achieve effective dimensionality reduction and robust performance across different datasets.

2026-07-15T00:00:00Z KHALDI, Brahim http://dspace.univ-mascara.dz:8080/jspui/handle/123456789/1441 2026-05-06T13:17:54Z 2026-05-06T00:00:00Z

Titre: Design of Bio-Inspired Metaheuristics for Medical Image Classification Auteur(s): KHALDI, Brahim Résumé: Breast cancer diagnosis relies fundamentally on histopathological examination of tissue samples; however, manual microscopic evaluation by pathologists is subjective, time-consuming, and prone to inter-observer variability. While deep learning has demonstrated remarkable success in medical image analysis, conventional single-architecture models exhibit inherent limitations. Convolutional neural networks excel at local feature extraction but fail to capture global contextual relationships, whereas Vision Transformers effectively model long-range dependencies but may be inefficient in representing fine-grained local details that are critical for histopathological interpretation. Moreover, optimizing deep learning models through hyperparameter tuning and feature selection remains computationally expensive and often suboptimal when relying solely on gradient-based optimization techniques. This thesis proposes HNet, a novel hybrid deep learning architecture that strategically integrates three complementary neural paradigms for histopathological image classification. HNet combines EfficientNet for efficient local feature extraction from tissue morphology, an Advanced Vision Transformer (AVT) for capturing global contextual relationships and long-range tissue patterns, and Capsule Networks (CapsNet) for explicitly modeling spatial hierarchies and part-whole relationships inherent in tissue organization. To address the challenges of high-dimensional feature spaces and suboptimal parameter tuning, a Genetic Algorithm (GA)-based feature selection mechanism is incorporated as a critical preprocessing step between the concatenated EfficientNet-AVT feature representations and the CapsNet input. This metaheuristic-driven optimization enables automatic identification of the most discriminative features while reducing dimensionality and computational complexity. Comprehensive experimental validation is conducted on the BreakHis dataset, encompassing binary (benign vs. malignant) breast cancer classification tasks. Detailed ablation studies quantify the individual contributions of each architectural component as well as the impact of GA-based feature selection on overall performance. Comparative evaluation against recent state-of-the-art methods demonstrates that the proposed HNet architecture achieves superior classification accuracy, sensitivity, specificity, and F1-score, establishing a new benchmark for breast cancer histopathology classification. The integration of metaheuristic-driven optimization with hybrid deep learning significantly enhances model robustness, generalization to unseen data, and computational efficiency compared to non- optimized approaches. Beyond empirical performance gains, this work demonstrates how hybrid metaheuristic, deep learning frameworks can be effectively integrated into clinical workflows by addressing real-world deployment constraints, including inference latency, computational resource utilization, and model interpretability. The proposed methodology provides a generalizable template for applying metaheuristic-optimized 5 hybrid deep learning to a wide range of medical image classification tasks, offering substantial potential for advancing automated computer-aided diagnosis systems and digital pathology.

2026-05-06T00:00:00Z ZAGANE, Mohammed http://dspace.univ-mascara.dz:8080/jspui/handle/123456789/1432 2026-05-03T10:20:34Z 2026-05-03T00:00:00Z

Titre: Algorithmics and Data Structures 1 : Lecture notes and solved exercises Auteur(s): ZAGANE, Mohammed

2026-05-03T00:00:00Z