Communications Internationales

Permanent URI for this collectionhttps://dspace.univ-boumerdes.dz/handle/123456789/11

Browse

Start date: 2020Subject: search.filters.subject.Machine Learning

Search Results

Now showing 1 - 5 of 5
  • No Thumbnail Available
    Item
    A Binary Relevance Approach for Smart Antenna Selection in Massive MIMO Systems
    (Institute of Electrical and Electronics Engineers, 2025) Bouchibane, Fatima Zohra; Boutellaa, Elhocine; Tayakout, Hakim; Cherigui, Rahma; Bouazabia, Sarah
    Future transceivers are projected to incorporate massive antenna arrays, which could significantly increase power consumption. To mitigate this challenge, the antenna selection technique (AS) emerges as a viable solution. By strategically selecting a subset of antennas, the system power consumption can be significantly reduced without compromising the overall system performance. This paper proposes a novel AS approach for massive MIMO systems under real-world channel measurements. By employing the binary relevance technique (BR), a straightforward approach to multi-label (ML) learning that tackles the problem by treating each class label as an independent binary classification task, we formulate the AS problem as a ML classification task. We conducted simulations using SVM as the base learning algorithm to assess the performance of our proposed approach and compare results to the Multi Label convolutional neural network (ML-CNN) and convex relaxation based approaches (CVX). The binary relevance based SVM (SVM-BR) performance, while slightly below the suboptimal convex relaxation approach in terms of system capacity, remains competitive with the MLCNN under different antenna array configurations
  • No Thumbnail Available
    Item
    Towards a Longitudinal Comparison Between Different Strategies for Android Malware Detection
    (Institute of Electrical and Electronics Engineers Inc, 2023) Mesbah, Abdelhak; Baddari, Ibtihel; Riahla, Mohamed Amine
    The growing popularity of the Android platform makes it a target of malware authors. The effective identification of such malware is an ongoing challenge. Several methods using machine learning have been proposed to prevent this threat. These methods are usually conventionally evaluated without considering the extent of performance over time. Given the evolving nature of both malware and benign apps, conventional evaluation may lack information. To imitate reality, this study compares the longitudinal performance of different machine learning models, using different strategies that combine permissions and API calls as features extracted through static analysis. Thus, to determine which strategy of features on which classifier are most effective to characterize malware for building a robust malware detector. To achieve this goal, on the one hand, we use a large real-world app set consisting of 100K (50k benign, 50k malware) apps date-labeled, collected across ten years, first seen between 2013 and 2022. On the other hand, each feature's strategy is fed into five classifiers (i.e., SVM, RF, LR, DT, and ANN), using old apps for the training and new apps for the evaluation. Among the assessed machine learning models, the SVM achieves the most promising results over time by employing the combination strategy of the high difference usage of API calls and permissions.
  • No Thumbnail Available
    Item
    Machine learning in the medical field: A comprehensive overview
    (Institute of Electrical and Electronics Engineers Inc, 2023) Belgacem, Ali; Khoudi, Asmaa; Boudane, Fatima; Berrichi, Ali
    Machine learning utilization in medicine has increased interest over the last few years. With its impressive results in treating diseases and medical conditions, it will be important to understand and analyze how the scientific community has used it. Thus, opening up space for new research and opportunities in medicine. The objective of this study is to review the literature on machine learning applications in the medical sector. Therefore, we conducted an extensive research by reviewing recent studies and surveys on machine-learning health solutions. As a result, we offer, in this paper, a fresh study affirming the foundations and necessities of a machine learning application in the medical field. We also provide a breakdown of current research trends, which highlights future research opportunities.
  • No Thumbnail Available
    Item
    ECG beats classification with interpretability
    (IEEE, 2022) Hammachi, Radhouane; Messaoudi, Noureddine; Belkacem, Samia
    Recently, a lot of emphasis has been placed on Artificial Intelligence (AI) and Machine Learning (ML) algorithms in medicine and the healthcare industry. Cardiovascular disease (CVD), is one of the most common causes of death globally, and Electrocardiogram (ECG) is the most widely used diagnostic tool to investigate this disease. However, the analysis of ECG signals is a very difficult process. Therefore, in this work, automated classification of ECG data into five different arrhythmia classes is proposed, based on MIT-BIH dataset. Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) Deep Learning (DL) models were used. The black-box nature of these complex models imposes the need to explain their outcomes. Hence, both Permutation Feature Importance (PFI) with Gradient-Weighted Class Activation Maps (Grad-CAM) interpretability techniques were investigated. Using the K-Fold cross-validation method, the models achieved an accuracy of 97.1% and 98.5% for CNN and LSTM, respectively
  • No Thumbnail Available
    Item
    Using Machine Learning for Heart Disease Prediction
    (Springer, 2021) Salhi, Dhai Eddine; Tari, Abdelkamel; Tahar Kechadi, M.
    In this paper we carried out research on heart disease from data analytics point of view. Prediction of heart disease is a very recent field as the data is becoming available. Other researchers have approached it with different techniques and methods. We used data analytics to detect and predict disease’s patients. Starting with a pre-processing phase, where we selected the most relevant features by the correlation matrix, then we applied three data analytics techniques (neural networks, SVM and KNN) on data sets of different sizes, in order to study the accuracy and stability of each of them. Found neural networks are easier to configure and obtain much good results (accuracy of 93%).