Papers by Sadok Ben Yahia
Arabic Aspect Category Detection for Hotel Reviews based on Data Augmentation and Classifier Chains
Proceedings of the 38th ACM/SIGAPP Symposium on Applied Computing
Multi-Label Learning for Aspect Category Detection of Arabic Hotel Reviews Using AraBERT
Proceedings of the 15th International Conference on Agents and Artificial Intelligence
Proceedings of the 14th International Joint Conference on e-Business and Telecommunications, 2017
Trust and reputation management stands as a corner stone within the Online Social Networks (OSNs)... more Trust and reputation management stands as a corner stone within the Online Social Networks (OSNs) since they ensure a healthy collaboration relationship among participants. Currently, most trust and reputation systems focus on evaluating the credibility of the users. The reputation systems in OSNs have as objective to help users to make difference between trustworthy and untrustworthy, and encourage honest users by rewarding them with high trust values. Computing reputation of one user within a network requires knowledge of trust degrees between the users. In this paper, we propose a new Clustering Reputation algorithm, called RepC, based on trusted network. This algorithm classifies the users of OSNs by their trust similarity such that most trustworthy users belong to the same cluster. We conduct extensive experiments on a real online social network dataset from Twitter. Experimental results show that our algorithm generates better results than do the pioneering approaches of the literature.
Revue Africaine de Recherche en Informatique et Mathématiques Appliquées, 2005
The extremely large number of association rules that can be drawn from ―even reasonably sized dat... more The extremely large number of association rules that can be drawn from ―even reasonably sized datasets―, bootstrapped the development of more acute techniques or methods to reduce the size of the reported rule sets. In this context, the battery of results provided by the Formal Concept Analysis (FCA) allowed to define "irreducible" nuclei of association rule subset better known as generic basis. However, a thorough overview of the literature shows that all the algorithms dedicated neglected an essential component: the relation of order, or the extraction of the minimal generators. In this paper, we introduce the GenAll algorithm to build a formal concept lattice, in which each formal concept is "decorated" by its minimal generators. The GenAll algorithm aims to extract generic bases of association rules. The main novelty in this algorithm is the use of refinement process to compute immediate successor lists to simultaneously determine the set of formal concepts, ...
This paper introduces a new personalized recommender system for folksonomies based on quadratic c... more This paper introduces a new personalized recommender system for folksonomies based on quadratic concepts.
IEEE/ACM Transactions on Computational Biology and Bioinformatics, 2018
Comparison of the protein interaction networks from different species is of paramount importance ... more Comparison of the protein interaction networks from different species is of paramount importance for understanding physical and functional interactions between biological functions and processes within a cell. In this paper, we introduce a novel algorithm for a global alignment of multiple proteinprotein interaction (PPI) networks called MAPPIN. The latter combines information available for the proteins in the networks, including sequence, function and network topology. Our method is perfectly designed to exploit current multi-core CPU architectures. MAPPIN has been extensively tested on a real dataset (five eukaryotic species). Our experimental results show that MAPPIN sharply outperforms the pioneering methods of the literature in producing functionally coherent alignments as far as it provides biologically significant alignments within an acceptable running time, even for very large input instances.
2012 IEEE 24th International Conference on Tools with Artificial Intelligence, 2012
Genetic Programming for Optimizing Fuzzy Gradual Pattern Discovery
Proceedings of the 7th conference of the European Society for Fuzzy Logic and Technology (EUSFLAT-2011), 2011
Artificial Intelligence and Soft Computing, 2010
Le nombre prohibitif de règles d'association qui peuvent être dérivées même pour des bases de don... more Le nombre prohibitif de règles d'association qui peuvent être dérivées même pour des bases de données de taille raisonnable est à l'origine du développement de techniques pour réduire la taille de l'ensemble de ces règles. Dans ce contexte, les résultats obtenus par l'Analyse de Concepts Formels (AFC) a permis de définir un sous-ensemble de règles appelé base générique. Cependant, un survol de la littérature montre que tous les algorithmes qui leur sont dédiés ont négligé une composante essentielle : soit la relation d'ordre sous-jacente ou c'est l'extraction des générateurs minimaux qui manque à l'appel. Dans ce papier, nous proposons un algorithme, appelé GENALL, pour construire un treillis de concepts formels, dans lequel chaque concept formel est "décoré" par ses générateurs minimaux. L'objectif de cet algorithme est d'extraire toute la connaissance nécessaire pour extraire les bases génériques des règles d'association, e.g ceux de Bastide et al. La principale originalité de GENALL est l'emploi d'un processus de raffinement des listes des successeurs immédiats pour déterminer, d'une manière simultanée, l'ensemble des concepts formels, leur ordre partiel sousjacent et l'ensemble de générateurs minimaux associés à chacun des concepts formels. Les résultats expérimentaux ont prouvé que l'algorithme proposé est particulièrement efficace pour des contextes d'extraction denses comparé à Nourine et al. Le temps de réponse de l'algorithme GENALL surpasse celui de l'algorithme de Nourine et al. (il est en moyenne 40 fois plus rapide que celui de Nourine et al.
Résumé. Dans la littérature, les travaux se sont principalement focalisés sur l’extraction des mo... more Résumé. Dans la littérature, les travaux se sont principalement focalisés sur l’extraction des motifs fréquents. Toutefois, récemment, la fouille des motifs rares s’est avérée intéressante puisque ces motifs permettent de véhiculer des connaissances concernant des événements rares, inattendus. Ils ont ainsi prouvé leur grande utilité dans plusieurs domaines d’application. Cependant, un constat important associé à l’extraction des motifs rares est d’une part leur nombre très élevé et d’autre part la qualité faible de plusieurs motifs extraits. Ces derniers peuvent en effet ne pas présenter des corrélations fortes entre les items les constituant. Afin de pallier ces inconvénients, nous proposons dans cet article d’intégrer la mesure de corrélation bond afin d’extraire seulement l’ensemble des motifs rares vérifiant cette mesure. Une caractérisation de l’ensemble résultant, des motifs corrélés rares, est alors réalisée en se basant sur l’étude des contraintes de nature différentes indu...
Multi-Label Learning for Aspect Category Detection of Arabic Hotel Reviews Using AraBERT
Proceedings of the 15th International Conference on Agents and Artificial Intelligence
Evolution d'ontologies: revue et critiques
HAL (Le Centre pour la Communication Scientifique Directe), 2010
Exploring OWL-DL Primitives for Automatic Ontology Alignment Tools Parametrization
ABSTRACT The increasing number of ontologies and the data that they encapsulate make the process ... more ABSTRACT The increasing number of ontologies and the data that they encapsulate make the process of ontology alignment an essential component of the Semantic Web by providing its inter operability. In recent years, several tools have been developed in order to produce alignments that reflect the correspondence degree between two ontologies to be aligned. The quality of alignments provided by these tools is closely related to certain parameters that govern their treatment. In this paper, we propose a new approach to automatically adjust the parameters of an ontology alignment method. This approach is based on the exploitation of OWL-DL primitives. It tends to establish in a dynamic way the weights assigned to different modules of an alignment method. The experiments show a marked improvement over a static and frozen previous settings.
Distributed Scalable Association Rule Mining over Covid-19 Data
Lecture Notes in Computer Science, 2021
Lecture Notes in Computer Science, 2019
Most of the information is available in the form of unstructured textual documents due to the gro... more Most of the information is available in the form of unstructured textual documents due to the growth of information sources (the Web for example). In this respect, to extract a set of events from texts written in natural language in the management change event, we have been introduced an open information extraction (OIE) system. For instance, in the management change event, a PERSON might be either the new coming person to the company or the leaving one. As a result, the Adaptive CRF approach (A-CRF) has shown good performance results. However, it requires a lot of expert intervention during the construction of classifiers, which is time consuming. To palpate such a downside, we introduce an approach that reduces the expert intervention during the relation extraction. Also, the named entity recognition and the reasoning, which are automatic and based on techniques of adaptation and correspondence, were implemented. Carried out experiments show the encouraging results of the main approaches of the literature.
A Neural-Based Approach for Extending OLAP to Prediction
Lecture Notes in Computer Science, 2012
In the Data Warehouse (DW) technology, On-line Analytical Processing (OLAP) is a good application... more In the Data Warehouse (DW) technology, On-line Analytical Processing (OLAP) is a good applications package that empowers decision makers to explore and navigate into a multidimensional structure of precomputed measures, which is referred to as a Data Cube. Though, OLAP is poorly equipped for forecasting and predicting empty measures of data cubes. Usually, empty measures translate inexistent facts in the DW and in most cases are a source of frustration for enterprise managements, especially when strategic decisions need to be taken. In the recent years, various studies have tried to add prediction capabilities to OLAP applications. For this purpose, generally, Data Mining and Machine Learning methods have been widely used to predict new measures' values in DWs. In this paper, we introduce a novel approach attempting to extend OLAP to a prediction application. Our approach operates in two main stages. The first one is a preprocessing one that makes use of the Principal Component Analysis (PCA) to reduce the dimensionality of the data cube and then generates ad hoc training sets. The second stage proposes a novel OLAP oriented architecture of Multilayer Perceptron Networks (MLP) that learns from each training set and comes out with predicted measures of inexistent facts. Carried out experiments demonstrate the effectiveness of our proposal and the performance of its predictive capabilities.
HAL (Le Centre pour la Communication Scientifique Directe), Mar 1, 2007
L'alignement d'ontologies revêt toute son importance dans des applications nécessitant la prise e... more L'alignement d'ontologies revêt toute son importance dans des applications nécessitant la prise en compte d'une interopérabilité sémantique. Plusieurs approches d'alignement d'ontologies existent dans la littérature. Elles sont basées sur les mesures de similarités. Dans ce papier, une nouvelle méthode d'alignement d'ontologies OWL-Lite est décrite. Le module d'alignement implémente une nouvelle approche d'alignement d'ontologies qui définit un modèle global de calcul de similarité, tout en remédiant au problème de l'intervention de l'utilisateur dans le processus d'alignement. Les tests expérimentaux réalisés sur les ontologies de benchmark montrent une nette amélioration des métriques de rappel et de précision. ABSTRACT. Ontologies have been established for knowledge sharing and are of extensive use as a means for conceptually structuring domains of interest. Thus, in order to guarantee a fluent global communication and knowledge exchange between local knowledge sketched by ontologies, the alignment of ontologies has emerged as a compelling topic to address. In this paper, we introduce a new approach for aligning OWL-Lite ontologies. The main originality of the alignment method stands in the fact that it palliates the main drawbacks appearing in the literature approaches, i.e., problem of user-parameter settings. Carried out experimental results pointed out a sharp improvement in the precision and recall evaluation metrics.
Herald of the Kazakh-British technical university
This study provides a detailed analysis and prediction of power generation at wind farms in Germa... more This study provides a detailed analysis and prediction of power generation at wind farms in Germany using Lasso, LightGBM, and CatBoost machine learning models. Feature Engineering was used on the data, which allowed the extraction of more detailed data, which was used to improve the quality of the models. Through Extensive Data Analysis (EDA), the authors identify and develop lagged and moving features from the energy production time series, under the assumption that accurate predictions can significantly improve the stability of energy systems, especially in the context of increasing dependence on renewable energy sources. The performance of each model is evaluated based on the Mean Absolute Error(MAE), Mean Squared Error(MSE), and Root Mean Squared Error(RMSE) metrics, with CatBoost exhibiting the highest accuracy. In conclude, pointing to opportunities for further research aimed at optimizing these models and adapting them to other regions, emphasizing the comprehensive and long...
Big Data Analytics in Association Rule Mining: A Systematic Literature Review
2021 the 3rd International Conference on Big Data Engineering and Technology (BDET), 2021
Due to the rapid impact of IT technology, data across the globe is growing exponentially as compa... more Due to the rapid impact of IT technology, data across the globe is growing exponentially as compared to the last decade. Therefore, the efficient analysis and application of big data require special technologies. The present study performs a systematic literature review to synthesize recent research on the applicability of big data analytics in association rule mining (ARM). Our research strategy identified 4797 scientific articles, 27 of which were identified as primary papers relevant to our research. We have extracted data from these papers to identify various technologies and algorithms of using big data in association rule mining and identified their limitations in regards to the big data categories (volume, velocity, variety, and veracity).
Uploads
Papers by Sadok Ben Yahia