• 15/12/2018
  • 22:45 (GMT +1)
My main research interests are described below. More details can be founded on website of my research group JARIR.

Knowledge Discovery refers to the overall process of discovering useful knowledge from data. In this process, data mining is an important step where we apply specific algorithms for extracting useful information (knowledge). There are various types of data mining techniques such as association rules, classifications and clustering. Association rule mining, one of the most important and well techniques of data mining. Algorithms for extracting association rules explore large item spaces to identify all items that satisfy some user-specified criteria. Then, these items are used to generate association rules. These algorithms suffer from two drawbacks: (i) the size of the search space to explore; (ii) the multiple accesses to physical databases. A known challenge (and always an open research issue) is how to constrain the search space to prevent an exponential explosion of rules while minimizing information loss. Our contributions to this challenge are twofold:
  • - Data representation and algorithmic : We study more suitable data to represent the search space in order to reduce its exploration. For example, we used intensively Formal Concept Analysis that can be used as an unsupervised clustering technique. In addition, FCA is particularly suited for exploratory data analysis because of its human-centeredness. The generation of knowledge is promoted by the FCA representation that makes the inherent logical structure of the information transparent. We used this representation for different types of data like text (text mining), web (web mining) and web services (web services discovery).
  • - Performance : To reduce the effect of size space on data mining algorithms performance, we proposed and developed parallel and distributed versions of sequential data mining algorithms (parallel and distributed generation of association rules, parallel and distributed classification, etc.).
Research Staff
NamePositionUniversityEmail
Yahya SlimaniFull Professor, Head of the groupISAMM Manouba, Tunisiayahya.slimani@fst.rnu.tn
Chiraz LatiriProfessor, Research team leaderISAMM Manouba, Tunisiachiraz.latiri3@gmail.com
Khedija ArourAssistant Professor, Research team leaderINSAT Tunis, Tunisiakhedija.arour@gmail.com
Ghada GasmiAssistant Professor, Research team leaderINSAT Tunis, Tunisiaghada.gasmi@gmail.com
Brahim BounhasAssistant Professor, Research team leaderISD Manouba, Tunisiabounhas.ibrahim@gmail.com
Hatem HaddadAssistant Professor, Research team leaderISAMM Manouba, Tunisiahaddad.hatem@gmail.com
Ines BouzouitaAssistant Professor, Research team leaderENIT Tunis El Manar, Tunisiaines.bouzouita@yahoo.fr
External Collaborators
NamePositionUniversityEmail
Samba NdiayeProfessorUCAD, Senegalsamba.ndiaye@ucad.edu.sn
Fode CamaraAssistant ProfessorUCAD, Senegalfode.camara@ucad.edu.sn
Mohamed El Hadi BenelhadjAssistant ProfessorUniversity of Constantine 2, Algeriabenelhadj@gmail.com

High Performance Computing (HPC) represents any computational activity requiring more than a single computer to execute a given task. As example, we can cite supercomputers and computer clusters that are used to solve advanced computation problems. HPC has also the capacity to handle and analyze massive amounts of data at high speed. In our research, we used HPC to explore three major issues:
  • - HPC for large and distributed data applications : Due to the exponential growth of data, today there is an ever increasing need to process and analyze big data. The huge size of the available data and their high-heterogeneity and high-dimensionality make large-scale data mining applications computationally very demanding. Moreover, the quality of the data mining results often depends directly on the amount of computing resources available. In fact, data mining applications are poised to become the dominant consumers of supercomputing in the near future. There is a necessity to develop effective parallel algorithms for various data mining techniques. However, designing such algorithms is challenging. In this research topic, we try to develop and experiment data mining algorithms that are able to benefit from the capabilities of HPC infrastructures.
  • - HPC programming : While discussions of HPC architectures have long centered on performance gains, that is not the only measure of success of HPC. The programmability of these architectures is as important as performance. Because the high heterogeneity of HPC architectures, their programmability concerns various HPC platforms: Multiprocessors, Cluster, Grid Computing, Multicores, etc. Today’s Petascale computing algorithms and applications are not structured to take advantage of these architectures. Recent studies estimate that 80-90% of applications use a single level of parallelism. Hence, it’s important to revisit old parallel programming models and languages to define parallel programming models for HPC. One approach is to use hybrid parallelism, i.e. parallelism at multiple levels with multiple programming models (MPI, OpenMP, MapReduce, Hadoop, CUDA, etc.), will be used in many applications. Hybrid parallelism may emerge because application speedup at each level can be multiplied by future architectures.
  • - Fault tolerance and HPC : With the emergence of Petascale systems, fault tolerance has received a lot of attention over the last decade. Most the existing fault tolerance techniques for parallel and distributed applications have been applied to high computing applications. Unfortunately, despite their relative success, existing approaches do not fit well with the challenging evolutions of large-scale systems. In a development of a fault tolerance technique for HPC architectures, scalability and performance are the most two important aspects that we need to take to our consideration. The goal of this research is to develop scalable fault tolerance techniques to help to make future high performance computing applications self-adaptive and fault survivable.
Research Staff
NamePositionUniversityEmail
Yahya SlimaniFull Professor, Head of the groupISAMM Manouba, Tunisiayahya.slimani@fst.rnu.tn
Moez Ben Haj HmidaAssistant Professor, Research team leaderENIT Tunis El Manar, Tunisiamoez.benhajhmida@gmail.com
Ghada GasmiAssistant Professor, Research team leaderINSAT Tunis, Tunisiaghada.gasmi@gmail.com
Riadh FrefitaAssistantESPRIT Tunis, Tunisiareadfrefita@gmail.com
External Collaborators
NamePositionUniversityEmail
Ibrahima NiangProfessorUCAD, Senegalibrahima_niang@hotmail.com
Hayat BendoukhaAssistant ProfessorUSTO University, Algeriabendoukhayat@gmail.com

Information retrieval (IR) is considered as the science of searching for relevant information including documents, images, video and other forms of media across databases, unstructured data and the World Wide Web. Nowadays information retrieval is widely known and used in the context of online web search engines. But Information has also many other fields of application domains, like biomedicine, social network, social science, genomic, geographic, etc. In this research topic, we study five research thematics: contextual IR, social IR, temporal IR, Arabic IR and testbeds for IR systems.
  • - Context-based retrieval : Information retrieval can benefit from contextual information to adapt the results to a user’s current situation and personal preferences. For example, performing the same query in different contexts often leads to different result rankings. Hence, semantics-based information retrieval is especially challenging because a change in context has an impact on the knowledge base content. In the context of information retrieval, the impact of a contextual aspect on the query results determines its relevance. In this issue, we study the problem of identification of relevant contextual information and his effect (or influence) on the performance of an information retrieval system. Our goal is to achieve high recall and precision for a specific user in a specific situation.
  • - Social-based retrieval : Recently the fields of information retrieval and social network analysis have contributed to the emergence of a new category of information retrieval systems, namely the SIR systems (Social Information Retrieval). These systems extend conventional IR to incorporate the social context of search and recommendation. A particular aspect of SIR is the modeling of the social interaction between people (social ties), which is used for enhancing recommendation systems. Our goal in this issue is to see how information retrieval systems can be enriched by analyzing these social ties, and particularly the strong and weak ties that measure the strength of relationship between people.
  • - Temporal information retrieval systems : Temporal Information Retrieval (TPR) is an emerging research area in the field of Information Retrieval. It is a fact of modern life that an enormous volume of information is created, exchanged, and stored electronically on the Web. Much of the content of stored resources is strongly time-dependent. Hence, an electronic resource can be identifying not only by his content but also by some temporal features, like creation or update date. These temporal features can be used to increase precision of search in an information retrieval system. In fact, classical IR techniques based on topic similarity alone are not sufficient for the search in temporal document collections. To overcome this limit, temporal dimension available in electronic resources (like documents) should be incorporated with document ranking for efficient retrieval. Our objective in this issue is twofold: (i) identifying the temporal characteristics of documents; (ii) incorporating these characteristics into information retrieval techniques in order to improve the retrieval effectiveness of an information retrieval system.
  • - Arabic information retrieval : Traditional information retrieval system was carried out essentially in English and fueled by the annual Text Retrieval Conferences (TREC) sponsored by NIST (the National Institute of Standards and Technology). NIST has accumulated large amounts of standard data (text collections, queries, and relevance judgments) so that IR researchers can compare their techniques on common data sets. More recently, IR researchers have found a real interest to study new languages other than English. Now, TREC includes multilingual data and other organizations sponsor similar annual evaluations for European languages (CLEF) and Asian languages (NTCIR) (Chinese, Japanese, and Korean). Arabic began to be included in the TREC cross-lingual track, and in the TDT (topic detection and tracking) evaluations. The availability of standard Arabic data sets from the NIST and the Linguistic Data Consortium (LDC) has in turned spurred a huge acceleration in progress in information retrieval and other natural language processing involving Arabic language. Arabic is an interesting case to study in IR, because Arabic is a highly inflected language. In this sub-topic, we study some problematic related to IR systems (lemmatization, morphological analysis, indexation) and we use the Hadith corpora as knowledge basis.
  • - Testbed for IR systems : The central problem of information retrieval (IR) is to find test collections. The existing data sets (like those provided by TREC) have played a key role to promote progress in the IR domain. However, given the significant increase of online content over the past few years and their diversity, and of the increasing rate of search queries, the current testbeds are either too small or not representative of the real applications of IR systems.
Research Staff
NamePositionUniversityEmail
Yahya SlimaniFull Professor, Head of the groupISAMM Manouba, Tunisiayahya.slimani@fst.rnu.tn
Chiraz LatiriProfessor, Research team leaderISAMM Manouba, Tunisiachiraz.latiri3@gmail.com
Brahim BounhasAssistant Professor, Research team leaderISD Manouba, Tunisiabounhas.ibrahim@gmail.com
Khedija ArourAssistant Professor, Research team leaderINSAT Tunis, Tunisiakhedija.arour@gmail.com
Faiza NajjarAssistant Professor, Research team leaderENSI Manouba, Tunisiafaiza.najjar@ensi.rnu.tn
Moez Ben Haj HmidaAssistant Professor, Research team leaderENIT Tunis El Manar, Tunisiamoez.benhajhmida@gmail.com
Hatem HaddadAssistant Professor, Research team leaderISAMM Manouba, Tunisiahaddad.hatem@gmail.com
Research Students, Associates and Assistants
NamePositionUniversityEmail
Ahlem BouziriAssistantISAMM Manouba, Tunisiaahlembou@yahoo.com
Saloua ZammaliPhD studentFST Tunis, Tunisiazammalisalwa@gmail.com
Wiem Fekih HassenPhD studentFSB Bizerte, Tunisiawiem.fekih@gmail.com
Malek HajjemPhD studentmalek.hajjem@gmail.com
Sourour Bel Haj RhoumaPhD studentsoullou@gmail.com
Soumaya GuesmiPhD studentsoumaya.g@live.fr
Meriem ZinglaPhD studentzinglameriem@gmail.com
Chedi BechikhPhD studentFSEG Jendouba, Tunisiachedi.bechikh@gmail.com
Wiem Ben RomdhanePhD studentbr.wiem@yahoo.fr
Nadia SoudaniPhD studentFST Tunis, Tunisianadia.soudani@gmail.com
Wiem LahbibPhD studentwiemlahbib88@hotmail.fr
Zeineb GhodbaniPhD studentFaculty of Gafsa, Tunisiaghodhbanizeineb@gmail.com
Marwa HammamiPhD studentmarwa.ch26@gmail.com
Mohamed EttalebPhD studentmouhamed.taleb@hotmail.fr
Mohamed ChbelPhD studentmohammedchebel@gmail.com
External Collaborators
NamePositionUniversityEmail
Ibrahima NiangProfessorUCAD, Senegalibrahima_niang@hotmail.com
Hayat BendoukhaAssistant ProfessorUSTO University, Algeriabendoukhayat@gmail.com
Kamel SmailiFull ProfessorUniversity of Lorraine, FranceKamel.Smaili@loria.fr
Lionel BrunieFull ProfessorINSA Lyon, FranceLionel.Brunie@insa-lyon.fr
Eric GaussierFull ProfessorIMAG Grenoble, Franceeric.gaussier@imag.fr
Catherine BerrutFull ProfessorIMAG Grenoble, FranceCatherine.Berrut@imag.fr
Lynda TamineFull ProfessorIRIT Toulouse, FranceLynda.Lechani@irit.fr
Amel BouzeghoubFull ProfessorTelecom Sud-Paris, FranceAmel.Bouzeghoub@it-sudparis.eu
Fabrice EvrardAssistant ProfessorENSEEIHT Toulouse, FranceFabrice.Evrard@enseeiht.fr
Bakhta AmraneAssistant ProfessorUniversity of Oran, Algeriafatema_amrane@yahoo.fr
Bilel El AyebAssistant ProfessorFSB Bizerte, Tunisiaelayeb_bilel@yahoo.fr

Grid and Cloud Computing has attracted researchers as an alternative to supercomputers for high performance computing. Grids enable access to shared computing power and storage capacity, while Clouds enable access to leased computing power and storage capacity. Although there is a difference in the fundamental concepts of Grid and Cloud computing that does not necessarily mean that they are mutually exclusive. For example, it is quite feasible to have a Cloud within a Grid, as it is possible to have a Grid as a part of a Cloud. We use these infrastructures to revisit some problems and challenges of classical parallel and distributed systems.
  • - Deployment of data mining algorithms on Grid’s : In this issue, we are interested by implementing and adapting parallel and distributed data mining algorithms on these architectures.
  • - Load balancing for data mining algorithms : in these infrastructures, some computational resources are heavily loaded while others have available processing capacity. In this issue, we study the problem of load balancing for data mining algorithms running on Grid platforms. In the context of Cloud’s, we use the load balancing to enhance the QoS.
  • - Fault tolerance for Grid’s : Fault tolerance is an important property for large scale computational grid systems, where geographically distributed nodes co-operate to execute tasks. Due to the number of components and their diversity, nodes, networks, disks and applications frequently fail, restart, disappear and behave unexpectedly. As the number of grid system components increases, the probability of failures in the grid computing environment becomes higher than that in a traditional parallel computing environment. Hence, support for the development of fault tolerant applications has been identified as one of the major technical challenges to address for the successful deployment of computational. Since the failure of resources affects job execution fatally, fault tolerance service is essential to satisfy QOS requirement in grid computing. Commonly utilized techniques for providing fault tolerance are job checkpointing, load balancing and replication. In case of complex scientific workflows where tasks can execute in well defined order reliability is another biggest challenge because of the unreliable nature of the grid resources. To study this problem, we proposed a dynamic colored graph for representing a Grid infrastructure. On the basis of this new representation, we defined and implemented a framework to tolerate failures of nodes and communication links. The same graphic representation is used to study the load balancing problem in Grid systems.
  • - Load balancing and QoS for Cloud computing : Cloud computing has grown in popularity in recent years thanks to technical and PAYG business model (Pay As You Go) benefits. In Cloud computing environments, application requirements can be characterized by quality of service (QoS) requirements such as availability, security, reliability, load balancing, etc., as mentioned in the Service Level Agreement (SLA). To improve the efficiency of the Cloud infrastructures, system, we suggest an approach to make load balancing more dynamic to better manage the QoS of multi-instance applications in the Clouds. This approach is based on limiting the number of requests that, at a given time, can be effectively sent and stored in queues of virtual machines through a queue load balancer for incoming user requests. Our simulation-based experimental results using production workload models show that the proposed provisioning technique detects changes in workload intensity (arrival pattern, resource demands) that occur over time and allocates multiple virtualized resources to satisfy QoS constraints.
Research Staff
NamePositionUniversityEmail
Yahya SlimaniFull Professor, Head of the groupISAMM Manouba, Tunisiayahya.slimani@fst.rnu.tn
Moez Ben Haj HmidaAssistant Professor, Research team leaderENIT Tunis El Manar, Tunisiamoez.benhajhmida@gmail.com
Research Students, Associates and Assistants
NamePositionUniversityEmail
Baya ChalabiPhD studentESI Alger, Algeriab_chalabi@esi.dz
Ameni MeskiniPhD studentISI Kef, Tunisiaameni.meskini@gmail.com
External Collaborators
NamePositionUniversityEmail
Belabass YagoubiFull ProfessorUniversity of Oran, Algeriabyagoubi@gmail.com
Ghalem BelalemFull ProfessorUniversity of Oran, Algeriaghalem1dz@gmail.com
Ibrahima NiangProfessorUCAD, Senegalibrahima_niang@hotmail.com
Mohamed RebbahAssistant ProfessorUniversity of Mascara, Algeriarebbah_med@yahoo.fr
Hayat BendoukhaAssistant ProfessorUSTO University, Algeriabendoukhayat@gmail.com
Mohamed Mahmoud Ould DeyeAssistantUCAD, Senegaldeyeother@yahoo.fr