Leila Hamdad

École Nationale Supérieure d'Informatique | ESI · Laboratory LCSI

Deep Learning (DL) techniques are effective for designing network intrusion detection systems (NIDS) but they lack leveraging IoT network topology. In the meanwhile, Graph Neural Networks (GNNs) consider both statistical properties and topological dependencies outperforming DL in complex IoT systems. However, three improvements are required: 1) Sca...

Federated Learning (FL) has emerged as a promising Machine Learning paradigm, enabling multiple users to collaboratively train a shared model while preserving their local data. To minimize computing and communication costs associated with parameter transfer, it is common practice in FL to select a subset of clients in each training round. This sele...

Association rule mining (ARM) is one of the most popular tasks in the field of data mining, very useful for decision-making. It is an NP-hard problem for which Genetic algorithms have been widely used. This is due to the obtained competitive results. However, their main drawback is the fitness computation which is time-consuming, especially when wo...

Precipitation nowcasting is very important to secure individuals and property against adverse events that may be triggered, as well as to optimize the management of water resources and so-called weather-sensitive economic activities. Achieving effective temporal and spatial nowcasting resolution poses additional challenges to the meteorological com...

Covid-19 pandemic has negatively impacted many areas, including the economy and health care facilities, and has left more than 5 million deaths worldwide. In this paper, we use functional data analysis methods to describe evolution of the number of cases and the number of deaths of Covid-19 in Africa. We perform functional principal component analy...

In this work, we are interested in the extraction of association rules on the data stream. We propose three strategies to deal with this problem: resolution by local vision, resolution by local and independent global vision, resolution by global vision. Since this task is very time consuming, the genetic algorithm for extraction of association rule...

In this work, kernel spatial relative risk function estimation is of interest. We consider the case where covariates that may affect the spatial patterns of disease are contaminated by measurement errors. Finite sample properties were carried out in order to illustrate our methodology with real cancer data. We perform relative risk functions estima...

Association rules mining (ARM) is an unsupervised learning task. It is used to generate significant and relevant association rules among items in a database. APRIORI and FP-GROWTH are the most popular and used algorithms nowadays for extracting such rules. They are exact methods that consist of two phases. First, frequent itemsets are generated. Th...

Artificial Intelligence (AI) algorithms follow probabilistic models instead of deterministic ones, therefore the results produced by the AI algorithms may fail occasionally. This represents a tremendous obstacle for large scale deployment of AI. Also, The centralized AI can lead to the possibility of data tempering and fraud. Moreover, the reliabil...

We propose a novel supervised classification algorithm for spatially dependent data, built as an extension of kernel discriminant analysis, that we named Spatial Kernel Discriminant Analysis (SKDA). Our algorithm is based on a kernel estimate of the spatial probability density function, which integrates a second kernel to take into account spatial...

User interest inference in social media is an important research topic with great value in modern personalization and advertisement systems. Using relationships characteristics such as strength may allow more refined inference. Indeed, due to influence and homophily phenomena, people maintaining strongest relationships tend to be and become more si...

Background Meta-heuristics are high-level methods widely used in different fields of applications. To enhance their performance, they are often combined to concepts borrowed from machine learning and statistics in order to improve the quality of solutions and/or reduce the response time.AimIn this paper, we investigate the use of feature selection...

In online social networks, the incomplete or noisy data are usual conditions raising increasingly the need for more accurate methods; especially in user attribute profiling. This work explores the influence of social tie strength in such settings, based on the intuition that the stronger the relationship is, the more likely its members are to share...

In mobile application stores, users very often rely on the opinions of others before downloading an application and its reputation could depend entirely on them. This makes analysis of users’ reviews very interesting for application owners to take future decisions. In this paper, we are interested in analyzing Algerian reviews on application store...

Several metaheuristics can be considered for solving a given optimization problem. Unfortunately none of them is better on all instances. Selecting a priori the best metaheuristic for a given instance is a difficult task which can be addressed using meta-learning. In this work, we propose a method to recommend, for a MaxSAT instance, the best metah...

With the emergence of social networking platforms and great amount of generated content, analyzing people interactions and behaviour raises new opportunities for several applications such as user interest profiling. In this context, this paper highlights the importance of considering relationship strength to infer more refined and relevant interest...

Sentiment Analysis and its applications have spread to many languages and domains. With regard to Arabic and its dialects, we witness an increasing interest simultaneously with increase of Arabic texts volume in social media. However, the Algerian dialect had received little attention, and even less in Latin script (Arabizi). In this paper, we prop...

We provide a tool to generate dataset for Visual Question Answering, based on Semantic analysis of existing dataset.

Developing a real-time VQA Framework based on ResNet50.

In this work, we investigate the use of unsupervised data mining techniques to speed up Bee Swarm Optimization metaheuristic (BSO). Knowledge is extracted dynamically during the search process in order to reduce the number of candidate solutions to be evaluated. One approach uses clustering (for grouping similar solutions) and evaluates only cluste...

Feature selection is often used before a data mining or a machine learning task in order to build more accurate models. It is considered as a hard optimization problem and metaheuristics give very satisfactory results for such problems. In this work, we propose a hybrid metaheuristic that integrates a reinforcement learning algorithm with Bee Swarm...

The first goal of this paper is to study the impact of Genetic Algorithms (GA’s) components such as encoding, different crossover and replacement strategies on the number and quality of extracted association rules. Moreover, we propose a strategy to manage the population. The later is organized in sub-populations where each one encloses same size r...

We study the asymptotic properties of the spectral density estimator (a periodogram) of a linear spatial process with alpha mixing innovations. A periodogram is a natural estimate of the spectral density. Under some conditions, a relation between the periodograms of innovations and that of the linear process is established in a spatial case. As the...

Choices of decision makers in a basketball team are not limited to the strategies to be adopted during games. The most important ones are outside the field and concern team composition and talented and productive players to acquire on which the team can rely to raise its game level. In this paper, we propose to use data mining tasks to help decisio...

The spatial data mining (SDM) is a process that extracts knowledge from large volumes of spatial data. It takes into account the spatial relationships between the data. To integrate these relations in the mining process, SDM uses two main approaches: Static approach that integrates spatial relationships in a preprocessing phase, and dynamic approac...

Spatial Data Mining allows users to extract implicit but valuable knowledge from spatial related data. Two main approaches have been used in the literature. The first one applies simple Data Mining algorithms after a spatial pre-processing step. While the second one consists of developing specific algorithms that considers the spatial relations ins...

Feature selection is an important data-preprocessing step that often precedes the classification task. Because of large amount of features in real world applications, feature selection is considered as a hard optimization problem. For such problems, metaheuristics have been shown to be a very promising solving approach. In this work, we propose to...

A nonparametric density estimate, that incorporates the spatial dependency, has not been studied in the literature. In this paper, we propose a new spatial density estimator which depends on two kernels: one controls the distance between observations while the other controls the spatial dependence structure. The uniform almost sure convergence of t...

A nonparametric density estimate, that incorporates the spatial dependency, has not been studied in the literature. In this paper, we propose a new spatial density estimator which depends on two kernels: one controls the distance between observations while the other controls the spatial dependence structure. The uniform almost sure convergence of t...

We propose in this work a new self organized biomimitic approach for unsupervised classification, named BFC, based on BBO (Biogeography based optimization). This method is tested on several real datasets(IRIS, Satimages and heart). These benchmarks are characterized by increasing overlap degree. Moreover, a comparison of BFC with other clustering m...

We investigate kernel estimates of the density and regression functions of linear processes with dependent innovations. The uniform almost sure and mean square consistencies of the estimates are proved under some mixing conditions. Special attention is paid to some simulations results.