Yahya Slimani

A survey on cloud service description

Cloud service description (CSD) becomes an active area which attracts the attention of many research organizations, while no standard CSD exists. This lack of standardization is caused by the vendor lock-in problem, where cloud providers use various techniques (languages, standards, ontologies, models, etc.) to describe cloud services. Furthermore, there are few studies which dealt with the four aspects of service especially the technical, operational, business, and semantic aspects. Whereas, many other studies dealt only with some of these four aspects. This is the case of WSDL (Web Service Description Language) language, which focuses in the technical aspect, but it does not cover business and semantic ones. Our objective in this paper is to identify the foundations of a standardized CSD covering the technical, operational, business, and semantic aspects. We present a comparative study of the different approaches that address CSD issues. The result of this comparative study shows that USDL (Unified Service Description Language) is the appropriate language supporting a service description by covering three aspects (technical, operational, and business), despite its inability to cover the semantic aspect.

A hybrid model for Arabic document indexing

Using either stems or roots as index terms offered considerable performance to Arabic Information Retrieval (IR) systems compared to the use of surface words for indexing. Many comparative works tried to find out the best from these two indexing approaches but until then, no of the two methods widely overtook the other. Each of the two index types performed better under different test circumstances in terms of recall and precision. In this paper, we propose a hybrid approach combining the two indexing units in a way we take the advantages from both of them and try to overcome their shortcomings. Then, based on some combining techniques, we assign a weight for each indexing unit and try to find out the best weighting values.

Combining Indexing Units for Arabic Information Retrieval

Using either stems or roots as index terms offered considerable performance to Arabic Information Retrieval IR systems compared to the use of surface words for indexing. Many comparative works tried to find out the best from these two indexing approaches but until then, no of the two methods widely overtook the other. Each of the two index types performed better under different test circumstances in terms of recall and precision. In this paper, the authors propose a hybrid approach combining the two indexing units in a way they take the advantages from both of them and try to overcome their shortcomings. Then, based on some combining techniques, the authors assign a weight for each indexing unit and try to find out the best weighting values.

Arabic Terminology Extraction and Enrichment Based on Domain-Specific Text Mining

Information Retrieval (IR) involves several disciplines to improve the search quality. In this paper, we focus on terminology extraction which includes critical tasks especially for highly ambiguous languages like Arabic. Properly retrieving a list of relevant terms for a given domain and enrich it is a persistent problem. Therefore, and in order to improve the coverage of terminology extraction and enrichment, we must strengthen our research by text-mining technologies based on well-founded methods. In this article, we present a new Arabic terminology extraction and enrichment approach which exploits the corpus structure as a first step to extract a minimal terminology and uses text-mining techniques to enrich it in a second step.

Distributed optimization of cyclic queries with parallel semijoins

We consider the problem of finding (possibly optimal) semijoin sequences that (fully) reduce the relations referenced in a cyclic query graph. We propose a combination of parallel and sequential semijoin operations to minimize the amount of data transmission in distributed query processing. We report on experiments that show that our approach is not only efficient but also effective in reducing the total amount of data transmission.

A Possibilistic Approach for Arabic Domain Terminology Extraction and Translation

This paper proposes a hybrid possibilistic approach for bilingual terminology extraction using possibility and necessity measures. On the one hand, we extract domain-relevant terms from the source language, and on the other hand, we build a co-occurrence-based translation graph, which is mined to translate terms in the target language. We compare our approach with different state-of-the art approaches. Experimental results show that the possibilistic approach reaches better results in terms of Recall, Precision and Mean Average Precision (MAP). The differences between the compared approaches show that our contribution is significant in terms of p-value.

Enhancing Hybrid Indexing for Arabic Information Retrieval

Existent literature proposes several approaches to enhance Arabic document retrieval using different indexing units. In anterior work [1, 2], we proposed to combine multiple indexing units which improved retrieval performance. This paper develops this approach and suggests enhancing term weighting through result aggregation and pseudo-relevance feedback techniques. We compare these approaches to three baselines to enhance the previous results which showed the performance of hybrid indexing. To assess our hypothesis, we run four experimental setups based on a larger corpus with various query sets. Finally, we aim to compare all these methods using standard information retrieval metrics.

Smart PDR integration for ubiquitous pedestrian navigation service

Quite recently, considerable attention has been paid to pedestrian navigation services, since they provide people with guidance and tracking information everywhere especially in an unfamiliar environment. In the literature, indoor and outdoor navigation has been widely investigated. However, studies on the continuity of pedestrian navigation service, still lacking. This paper presents a novel positioning technology selection algorithm able to integrate the pedestrian dead reckoning (PDR) technique, for short-term, through the navigation process either to improve the calculation of the mobile units (MU) position and/or to maintain the continuity of the service. The integration of the PDR is inferred based on the fuzzy logic approach. In fact, the proposed algorithm defines a fuzzy variable called “margin” which identifies four different states, namely, confidence, intermediate, confusion and critic. The PDR technique is only incorporated in the last two states in two ways: either fused with the used radio frequency (RF) positioning technology or stand-alone if no RF technology is available. The algorithm performance is evaluated based on an empirical study.

Towards a standardized cloud service description based on USDL

Abstract In recent years, cloud computing paradigm has attracted a lot of attention from both industry and academia. However, each cloud provider uses its own techniques (languages, standards, ontologies, or models, etc.) to describe cloud services. The diversity of these techniques leads to the vendor lock-in problem, and thus, the lack of a cloud service description standardization. In addition, existing service descriptions cover only particular aspects and neglect others. For example, WSDL covers only technical aspect and does not cover business and semantic ones. Our objective is to define a standardized cloud service description that covers technical, operational, business, and semantic aspects. In this paper, we introduce different approaches that have dealt with cloud service description, and thus, we adopt USDL language as an appropriate technique to describe cloud services thanks to its expressivity by covering three perspectives (technical, operational, and business). But, USDL is still limited because it cannot cover semantic aspect and it is not intended for cloud computing domain. After that, we highlight USDL limitations that can appear in cloud computing domain and that should be taken into consideration in our research work. This paper will focus on establishing a WSMO-based ontology to define semantically cloud services. This new cloud service description is based on USDL and we will enhance it by taking into consideration some USDL limitations. Finally, we test our proposed cloud service description model on a case study to prove its applicability.