Faculté des mathématiques et de l'informatique

Permanent URI for this communityhdl:123456789/17

Browse

Search Results

Now showing 1 - 10 of 439
  • Thumbnail Image
    Item
    Etude Comparative Des Racinisateurs Arabes.
    (university of bordj bou arreridj, 2024) MAOUCHE Farah Hiba; KHALFA Lynda
    Arabic language has a complex morphological structure, which presents a unique challenge in natural language processing (NLP). The derivational system of Arabic is based on roots, which are frequently modified to create new words, employing an extensive set of Arabic morphemes affixes such as prefixes, suffixes and more. Stemming is a fundamental task in text processing, plays a crucial role in information retrieval (IR) and text analysis, it reduces words to their basic or root form, facilitating text normalization for easier processing. However, no stemming algorithm for this language is perfect. In this work, we are going to focus on comparing and evaluating the performance of several Arabic stemmers namely, ISRI, Tashaphyne and Snowball. We intend to assess their performance on two distinct datasets using advanced techniques such as neural networks and machine learning classifiers. Additionally, we aim to determine which combination of stemmer and classifier yields the best results, providing invaluable insights for Arabic text processing applications
  • Thumbnail Image
    Item
    Deep Learning Algorithms for Remote Sensing
    (university of bordj bou arreridj, 2024) Bouketir Hadda; Habbeche Dounia
    Deep learning has revolutionized the analysis of data collected from unmanned aerial vehicle (UAV) imagery, allowing for more profound insights, precise analysis, and enhanced data extraction. This advancement has significantly contributed to the refinement of semantic segmentation techniques. Particularly convolutional neural networks (CNNs) have emerged as powerful tools in this domain, outperforming traditional methodologies. Nonetheless, challenges persist, including feature extraction, class imbalance issues, overfitting, and vanishing gradients that hinder deep neural network training, consequently impacting segmentation performance. To address these challenges, we propose a novel approach by integrating the U-Net architecture with ResNet34 backbone leveraging its strong feature extraction capabilities. These features are further improved by using trained weights from the ImageNet dataset. We train and evaluate the proposed model on several UAV datasets, including Aerial Semantic Segmentation, LandCover.ai, UAVid, and AeroScapes. We achieve remarkable performance, higher accuracy, precision, recall, F1-score, and miou compared to other methods.
  • Thumbnail Image
    Item
    ADMINISTRATION DES SYSTEMES D’INFORMATION
    (2025) Dr. Hakima ZOUAOUI
  • Thumbnail Image
    Item
    Classification automatique de la maladie de Parkinson à partir de la voix
    (university of bordj bou arreridj, 2024)  Dahili Ahlem; Benmessahel Samira
    Speech analysis is a promising approach for early and automated diagnosis of Parkinson's disease. This non-invasive and inexpensive method relies on the characteristic voice changes of the disease, present from the early stages, to identify patients. Automated systems based on artificial intelligence can analyze these voice changes and effectively discriminate Parkinson's disease patients from healthy subjects. Despite challenges such as voice variability and background noise, speech analysis has great potential to improve the diagnosis and management of Parkinson's disease. Ongoing research aims to refine this technology and make it a valuable tool for improving the quality of life for patients.
  • Thumbnail Image
    Item
    Suivre les rumeurs dans les réseaux sociaux
    (2024) LAALAOUI Mouna; LAHRI Sarah
    The emergence of the Internet has transformed global communications, making information accessible at unprecedented speeds. With the emergence of social media, people can now connect, share and interact immediately with a variety of content. However, the ease with which information can be exchanged also facilitates the rapid spread of unverified rumors and false information. In this work, we aim to track and detect rumors, and to this end, we will present a model based on a deep learning approach using LSTM and RNN algorithms in order to obtain the best possible classification and more accurate and valid results.
  • Thumbnail Image
    Item
    Amélioration des profits des agricultures dans les système de rotation cultural avec des contraintes d’adjacence des parcelles. Etude de cas : Daira de Medjana.
    (2023) Maouche Fadia; Chetioui Wiame
    The purpose of this project is to propose and analyze an optimization model for crop rotation on several surfaces focused on the agricultural production in the Daïra of Medjana. The objective is to optimize the use of land subject to neighborhood and succession constraints for crops of the same botanic family, in addition of course to maximizing the farmer’s pro t. This project introduces the basic concepts of crop rotation and agricultural planning, presents a model which takes the form of a linear program -implemented on Matlab- to determine better crop planning on a set of plots.
  • Thumbnail Image
    Item
    Using Multi-objective Meta-heuristics for Data Mining
    (university of bordj bou arreridj, 2025-01-05) Soumaia KAHLOUL
    The ability to extract knowledge from large datasets is essential for innovation and informed decision-making, a process known as knowledge extraction or data mining. Traditional methods often fall short in fully utilizing data potential, necessitating the development of new algorithms for better insights. This thesis explores an innovative approach by integrating deep learning with advanced feature selection techniques to improve the classification accuracy of COVID-19 cases from chest X-ray images. The dataset includes X-ray images categorized as COVID-19, pneumonia, and normal. We employ the Binary Multi-Objective Henry Gas Solubility Optimization Algorithm (B-MOHGSO) for feature selection and leverage models like AlexNet, VGG19, GoogleNet, and ResNet for feature extraction. Eight versions of B-MOHGSO were tested, with k-nearest neighbors (k-NN) as the classifier. The study highlights the significant impact of S-shaped and V-shaped transfer functions on binary transformations and classifier performance in high-dimensional medical imaging. Notably, B-MOHGSO algorithms, particularly those using V-shaped transfer functions, excelled in selecting relevant features while maintaining high accuracy. When combined with the VGG19 model and SVM classifier, B-MOHGSO significantly reduced the feature set without sacrificing performance. The application of B-MOHGSO in COVID-19 classification is crucial for identifying key features that enhance diagnostic processes and treatment strategies. By adapting MOHGSO for discrete optimization, this research aims to address the complexities of high-dimensional medical data and improve healthcare analytics outcomes.
  • Thumbnail Image
    Item
    Autism Detection Using Machine Learning
    (2024-09-23) Youcef Kouadria; Sakhraoui Zineddine
    This research investigates the use of machine learning, specifically deep learning, for the detection of Autism Spectrum Disorder (ASD) using facial image data. By employing the VGG16 model with transfer learning, the project achieved high classification accuracy, demonstrating the potential of machine learning in supporting early ASD diagnosis. The use of automated image analysis provides a non-invasive, scalable solution that could complement traditional diagnostic methods. The results indicate that machine learning can significantly contribute to healthcare by enabling quicker and more accurate ASD diagnoses. Future improvements could include the use of larger datasets and multimodal data such as voice and behavioral analysis
  • Thumbnail Image
    Item
    Towards biometric recognition system based on explainable classifier methods
    (Université de Bordj Bou Arreridj Faculty of Mathematics and Computer Science, 2024) Djouamai, Zineb
    er the past decade, biometric systems have advanced significantly, achieving high classifica tion accuracy and minimal equal error rates (EER). However, many conventional methods lack transparency and explainability, which are critical in areas like security and identity verifica tion, where trust is paramount. This limitation restricts the ability to understand these systems’ decision-making processes, making it difficult to ensure reliability and accountability in sensi tive applications. To address these challenges, we propose the development of an efficient biometric system based on explainable, rule-based classifiers. Unlike traditional approaches, our method incorpo rates explainability at its core, offering clear insights into the system’s decision-making process while maintaining high performance. This approach ensures that the system is not only accurate but also adaptable and user-friendly, enabling its application across a range of classification and predictive tasks. By prioritizing transparency alongside performance, the proposed system aims to meet the growing demand for trust and usability in biometric applications. Its dual focus on achieving low EER and delivering explainable outcomes ensures it is suitable for deployment in critical domains. This balance between accuracy and explainability positions the system as a reliable and advanced solution for high-stakes environments like security and identity management
  • Thumbnail Image
    Item
    outils de la programmation mathématique ( Maple)
    (Université de Bordj Bou Arreridj Faculty of Mathematics and Computer Science, 2024) Attia, Abdelouahab