WEBI 2012 Abstracts


Full Papers
Paper Nr: 2
Title:

Web Services based Approach for Integrating Multimedia Content

Authors:

Fatma Zohra Lebib, Hakima Mellah and Youssef Amghar

Abstract: Multimedia content is derived from various autonomous, distributed and heterogeneous content sources. To address problems posed by content sources heterogeneity, a service-oriented architecture is proposed to assure a dynamic integration of multimedia content. In this work we propose a model that we call MaaS (for Multimedia as a Service), through which multimedia content providers expose their content. Once the MaaS is discovered, its classification into category of concepts, based on domain ontology, is made. However, content access is done through a concept hierarchy. The sport domain is used to validate the proposed model.
Download

Paper Nr: 3
Title:

Matching Resources in Social Environment

Authors:

Amel Benna, Hakima Mellah, Islam Choui and Ali Oualid

Abstract: User comments on the web are becoming more and more important. We focus, in this paper, on the use of user-defined tags for annotating resources to identify links between them. These links are based on a social context of the resource, obtained by applying k-means classification method and a hierarchical classification of tags within a cluster. The resources are re-assigned to this classification to facilitate the search process. The ranking of results is performed according to their degree of relevance, by evaluating a similarity score between the tagged contents, in hierarchical clusters of tags, and the user request. The results of the evaluation, on the social bookmarking system del.icio.us, demonstrate significant improvements over traditional approaches.
Download

Paper Nr: 5
Title:

Web Usage Mining for Automatic Link Generation

Authors:

Olatz Arbelaitz, Ibai Gurrutxaga, Aizea Lojo, Javier Muguerza, Jesús M. Pérez and Iñigo Perona

Abstract: During the last decades, the information in the web has increased drastically but larger quantities of data do not provide perse added value for web visitors; there is a need for more efficient access to the required information and adaptation to user preferences or needs. The use of machine learning techniques to build user profiles allows to take into account users’ real preferences. We present in this work a preliminary system, based on the collaborative filtering approach, to identify and generate interesting links for the users while they are navigating. The system uses only web navigation logs stored in any web server (according to the Common Log Format) and extracts information from them combining unsupervised and supervised classification techniques and frequent pattern mining techniques. It also includes a generalization procedure in the data preprocessing phase and in this work we analyze its effect on the final performance of the whole system. We also analyze the effect of the cold start (0 day problem) in the proposed system. The experiments show that the proposed generalization option improves the results of the designed system, which performs efficiently w.r.t. a web-accessible database and is even able to deal with the cold start problem.
Download

Paper Nr: 6
Title:

Elicitation of Specific Requirements of Data Quality during the Web Portal Development

Authors:

César Guerra-García, Ismael Caballero, Rodrigo Testillano, Rafael Llamas and Mario Piattini

Abstract: Data is one of the most important assets for making decisions and concretizing business in organizations. So, providing data with adequate levels of quality, especially for Internet applications is a very important issue. However, most of developers of these applications do not take in account the incorporation of artifacts to the necessary management of data quality (DQ) from the early stage of development. Due to that, we have elaborated a strategy with two approaches: methodologic and technologic. The first one is aimed to identification the requirements corresponding to the Web portal functionalities for the different kind of users and their specific DQ software requirements. For the technologic approach, an UML profile to model the DQ software requirements is shown; it embraces aspects considered basics to integrate in the specification and modeling of these kinds of requirements. The final objective is developing applications that satisfy the different DQ software requirements specified by each user, at the moment to perform a function with the system.
Download

Paper Nr: 7
Title:

Expertise Search in Unstructured Data in ECM using S-BPM Approach

Authors:

Alexander Gromoff, Julia Stavenko, Kristina Evina and Nikolay Kazantsev

Abstract: This article describes the application of currently most promising methods of graph theory, content analysis and (3) subject-oriented approach to business process modelling for creating and automation of innovative process and therefore for maximization of ROI (return on investments) in intellectual and social capital of enterprises. In a course of development, instant full-text indexation takes place and taxonomic picture of different branches for such community is formed. In due course system gathers the statistics and builds-up maps of intercommunication with priority allocation of most discussed topics. A group of predetermined experts begins discussion on development prospects of this or that subject afterwards. The strategic map of investments into innovative development that can be offered to group of investors for competitive investments eventually turns out. In this process all steps except final (gathering of experts) are human non-dependant, what increase efficiency of the process in general.
Download

Paper Nr: 8
Title:

Using Social Networks based on Data Mining to Promote New Technologies for Women from Emerging Countries

Authors:

Syrine Ben Meskina, Valérie Monfort and Achraf Ben Miled

Abstract: Nowadays Social Network Websites become increasingly numerous and millions of people are registered. They become famous and useful in several domains, in particular for sales and marketing activities all over the world. Their success is gradually growing in industrialized countries. In our paper, we firstly show how this technology is used to promote and enlarge sales networks. Secondly, we investigate the ontologies dedicated to represent the Social Network Semantics and their limits. We propose to use this approach to a research project which aims at addressing women from both rural and urban areas of emerging low income countries, who are individually or within group involved with sustainable hand craft and/or fair trade business leading to a varying degree of economic mobility. We finally conclude by proposing a solution that aims to promote Sales Social Networks based on Data Mining techniques.
Download