MAKHLOUF Benazi

Aya HACHADI

Land Registry Traceability System Based on Blockchain Technology

MEBRAK MAHDI , ZAKKAR CHEYMA

All police : une plateforme pour délivrer des Amendes

LAHMAR KHAWLA , LAMDJAD DOUAA

Développement d’un système de recommandation pédagogique

Clustering is essential for identifying patterns in data by grouping similar points. However, many advanced algorithms face challenges when dealing with clusters of varying shapes and sizes. In this paper, we propose a KNN-FN hybrid algorithm combines the strengths of K-Nearest Neighbors (KNN) and the Fast Newman (FN) community detection algorithm to enhance clustering performance. KNN is used to construct a graph that captures local neighborhood structures by connecting each data point to its nearest neighbors, while the FN algorithm applies modularity maximization to detect well- defined clusters within the graph. This hybrid approach improves clustering, particularly in complex datasets with irregular shapes and varying densities, The KNN-FN hybrid algorithm efficiently detects clusters in large-scale data, making it suitable for real-world applications.

Citation

Nassira LOGRADA , Makhlouf Benazi , ,(2025-11-25), A Graph-Based Hybrid Clustering Approach for Detecting Complex Structures,The 2nd International Workshop on Machine Learning and Deep Learning (WMLDL25),msila

This paper presents an enhanced version of the DBSCAN algorithm, termed myDBSCAN, which simplifies the clustering process by relying on a single parameter, `eps`, instead of the traditional two parameters used in the original DBSCAN. The study evaluates myDBSCAN's performance across various synthetic datasets with different shapes and levels of noise. Empirical results demonstrate myDBSCAN’s performance comparably to the original DBSCAN algorithm, successfully identifying clusters with similar accuracy. The simplicity of using only one parameter makes myDBSCAN more accessible and easier to implement. The results highlighted myDBSCAN's effectiveness in clustering tasks, offering a practical alternative to traditional DBSCAN, particularly in scenarios where ease of parameter tuning is crucial. Future research will explore further optimizations and applications of myDBSCAN to a broader range of datasets.

Citation

Makhlouf Benazi , ,(2024-12-10), Simplifying the DBSCAN Algorithm with a Single-Parameter Approach.,The Sixth International Symposium on Informatics and Its Applications (ISIA),University of M’sila

The rapid development of the internet and social network platforms has given rise to a new field of research, social network analysis. This field of research has many fundamental problems, one of which is community detection. The objective of this research is to understand hidden connections among individuals. However, uncovering these connections are still challenging, despite the existence of several methods. In this paper, we propose a new algorithm called MCCD (Modified Cosine for Community Detection) for community detection in social networks based on node similarity. Our algorithm consists of two steps. In the first step, we use a novel cosine similarity formula to identify initial communities. In the second step, we merge these communities based on a new similarity measure. MCCD can be used in two different ways. The first way uses K as an input to identify the exact communities. The second way does not require K and aims to provide the best partitioning by maximizing modularity. Our algorithm has been tested on a variety of artificial and real-world networks, and the experimental results demonstrate its superiority over existing methods in detecting communities.

Citation

Makhlouf Benazi , , (2024-07-03), A robust two‑step algorithm for community detection based on node similarity, The Journal of Supercomputing, Vol:80, Issue:, pages:23592–23608, Springer Nature

Density-Based Spatial Clustering of Applications with Noise (DBSCAN) is one of the most important data clustering algorithms. Its importance lies in the fact that it can recognize clusters of arbitrary shapes and is not affected by noise in the data. To identify clusters, DBSCAN needs to specify two parameters: the parameter Eps, representing the radius of the circle to identify the neighborhood of each observation. The second parameter of DBSCAN is minpts, which represents the minimum size of the neighborhood for a point to be a seed in a cluster and not a noise. However, the task of determining the adequate value of Eps parameter is not easy and represents a major issue when applying DBSCAN since the accuracy of this algorithm highly depends on the values of its parameters. In this paper, we present a new version of DBSCAN where we need only to specify the minpts parameter, then we use k-nearest neighbors (kNN) algorithm to calculate the value of Eps automatically for every point in the data. This technique not only reduces the number of parameters by eliminating Eps which has been very difficult to determine, but also gives DBSCAN the ability to detect clusters with heterogeneous density. The experimental results show that the proposed method is more efficient and more accurate than the original DBSCAN algorithm.

Citation

Makhlouf Benazi , , (2024-04-18), An Adaptative Eps Parameter of DBSCAN Algorithm for Identifying Clusters with Heterogeneous Density, Computación y Sistemas, Vol:28, Issue:2, pages:465–472, https://www.cys.cic.ipn.mx/ojs/index.php/CyS/about

The community structure is proving to have a very important role in the understanding of complex networks, but discovering them remains a very diÕcult problem despite the existence of several methods. In this article, we propose a novel algorithm for discovering communities in complex networks based on a modiÒed random walk (RW) and label propagation algorithm (LPA). First, we calculate the similarity between nodes based on the new formula of RW. Then, the labels are propagated by the obtained similarity of the Òrst step using LPA. Finally, the third step will be a new measure to Ònd the optimal partitioning of communities. Experimental results obtained on several real and synthetic networks reveal that our algorithm outperforms existing methods in Ònding communities.

Citation

Makhlouf Benazi , BILAL Lounnas , Rabah Mokhtari , , (2022), A complex network community detection algorithm based on random walk and label propagation, Transactions on Emerging Telecommunications Technologies., Vol:33, Issue:9, pages:73-91, Wiley