Vol.13 No.1&2 March 1, 2014

Research Articles:

Enhanced LBS Discovery in a Decentralized Registry Based Web Services Environment (pp001-023)
        Melwyn D’Souza and V.S. Ananthanarayana
Location Based Services (LBS) is the most happening thing in the mobile industry today. Everybody is trying to generate revenue from location based services. Mobile phone manufacturers are developing new smart phones every day and network providers are offering high speed data connections. Several LBS providers and applications are available in the market but the major problem is service provider dependency. This paper gives an overview of a decentralized registry based architecture using web services technology which facilitates dynamic discovery, interoperability and provider independence. The web services technology uses UDDI registry service for publishing and discovering services but the discovery results obtained are not reliable as the service discovery considers only static service description. This paper contributes to enhancing LBS discovery by considering service dynamics and expanding LBS discovery process to neighboring locations.

Query Intent Detection Based on Query Log Mining (pp024-052)
        Juan Zamora, Marcelo Mendoza, and Hector Allende
In this paper we deal with the problem of automatic detection of query intent in search engines. We studied features that have shown good performance in the state-of-the-art, combined with novel features extracted from click-through data. We show that the combination of these features gives good precision results. In a second stage, four text-based classifiers were studied to test the usefulness of text-based features. With a low rate of false positives (less than 10~\%) the proposed classifiers can detect query intent in over 90\% of the evaluation instances. However due to a notorious unbalance in the classes, the proposed classifiers show poor results to detect \textbf{transactional} intents. We address this problem by including a cost sensitive learning strategy, allowing to solve the skewed data distribution. Finally, we explore the use of classifier ensembles which allow to us to achieve the best performance for the task.

Processing Mutliple Requests to Construct Skyline Composite Services (pp053-066)
        Shiting Wen, Qing Li, Chaoguang Tang, An Liu, Liusheng Huang, and Yangguang Liu
The performance of a composite service is determined by the performance of involved component services. When multiple non-functional criteria are considered, users are required to express their preferences over different quality attributes as numeric weights in existing methods. However, this imprecise method may not reflect the natural ordering of services and thus could miss some user-desired services. In this paper, we propose a composition framework to construct multiple skyline composite services for each individual request. We also discuss how a service registry can effectively deal with multiple requests simultaneously by materializing the intermediate composite services. We evaluate the efficiency and effectiveness of our methods through extensive experiments.

Web Application for an Adaptive Multi-Agent e-Learning System: a Continuous Improvement of e-Content (pp067-086)
        Salah Hammami and Hassan Mathkour
This paper presents a methodology for the continuous improvement of e-content in an e-learning system based on agent technology. Multi-agent systems play an important role in today's software development of web application. For this purpose, we introduce here a novel concept involving the development of an adaptive e-Learning web application that explores several recent technologies and including web design. The Adaptive E-Learning system based on agent technology is useful in developing continuous improvement strategies. The goal this work is to propose a generic model to assess and evaluate the students learning outcomes related to e-content. Based on the analysis and evaluation of the students learning outcomes results, the instructor identifies the low achievers of the e-content and makes a correction plan to improve them. An exploratory implementation has been developed and used in practice. database.

C1: an Automated Online Eduication Management System Based on an Object-Oriented Approach (pp087-096)
        S.C. Ng, T.S. Lee, C.K. Wong, and F.Y. Lee
Communication and information are integral parts of the education process and thus the integration of Information and Communication Technologies (ICT) in education management systems comes naturally. Many business organisations, schools and universities are gradually increasing the use of ICT to reduce costs and improve the efficiency of administration. This paper presents the implementation of an automated online education management system, named Cybernetics 1 (C1). C1 was tested as an operational tool to support the education process of undergraduate students. The four major modules in C1 include the registration of new students (inclusive of retrieving and updating of student profile), creation of module assessment structures, module marks entry and generation of academic transcripts. C1 was developed based on the object-oriented approach. This system successfully eliminates the problem of poor data sharing and data inconsistency as encountered by most existing education management systems. In particular, the normalisation process has significantly reduced the problems of data redundancy in the database.

A Fast Approach to Querying Multiple Ontology Versions Based on Concept Lattice (pp097-113)
        Yaqing Liu, Rong Chen, Yingjie Song, and Wu Deng
In this paper we propose a fast approach for querying multiple ontology versions, in which a novel model named as a version lattice based on a concept lattice is developed to serve as its foundation. We depict formally related problems of multiple ontology versions query and prove the equality between our description and a logic-based one. Supporting various requirements on multiple ontology versions, our approach can save more running time than previous algorithms, which is explained by analyzing our algorithms. Also experiments show that the advantage of running speed is more remarkable for ontology versions which are large in quantity and scale.

Empirical Study of Load Time Factor in Search Engine Ranking (pp114-128)
        Jakub Marszałkowski, Jedrzej M. Marszałkowski and Mackej Drozdowski
Search engine ranking position is essential for marketing plans of many e-businesses. There is a lot of confusion over factors influencing this position. One of them is time performance factor. The purpose of this paper is to identify the role of load time factor in the algorithm ranking the search results in Google. We present empirical study of 40 phrases and load times measured for the first 30 results for each phrase (1200 websites). Two types of load time factors used by Google were analyzed. To quantitatively confirm the results, simulation of the algorithm is proposed. The simulations show that the load time factor plays role in the Google algorithm, although its weight is not high. From two studied metrics Google seems to use the crawl time, i.e. time spent downloading a page by Google robot. A second set of experiments confirms that load time is not only correlated with the ranking position, but also effectively determines this position.

Towards a Homogeneous Characterization of the Model-Driven Web Development Methodologies (pp129-159)
        F.J. Domínguez-Mayo, M.J. Escalona, M. Mejías, M. Ross and G. Staples
In recent years a large number of Model-Driven Web development approaches have been designed and are being applied with success in real environments. However, as new ones are frequently emerging in this changing time, authors have to change and update them constantly and, consequently; development teams do not know which is the most suitable for them because in many cases it depends on their project scope. Furthermore, approaches are usually appearing with different concepts and terminologies in many cases, although all lack the use of standards and practical experience. Thus, the need of managing quality in this type of approach arises every day. This paper suggests a characterization of these methodologies in order to use this information for the quality management of Model-Driven Web development methodologies for authors and development teams alike. In addition, an experimental study in order to analyse and evaluate a Model-Driven Web development methodology (the NDT methodology) has been carried out within a specific work context.

ACOTA: A Multilingual and Semi-Automatic Collaborative Tagging Web-based Approach (pp160-180)
        Cesar Luis Alvargonzález, Jose Maria Álvarez Rodríguez, Jose Emilio Labra Gay, and Patrcia Ordoñez de Pablos
This paper introduces a multilingual hybrid methodology to automatically deploy and combine collaborative tagging techniques based on user-behavior and recommendation algorithms. A reference web architecture called ACOTA (Automatic Collaborative Tagging) is also described in order to show the recommendation capabilities of this approach with the aim to assist users when multilingual resource tagging is required. Finally a quantitative research in the context of corporate knowledge management is also presented to evaluate and assess the goodness and accuracy of the methodology to minimize the effort of multilingual document categorization.

Back to JWE Online Front Page