Catálogo de publicaciones - libros

This paper presents a dynamic and scalable mechanism for discovery of semantically enriched descriptions of Web services. By employing Web Service Modeling Ontology (WSMO) as the underlying framework for describing both user requests and Web services, and combining it with the usage of Peer-to-Peer technology in this context, a scalable, distributed, dynamic and flexible discovery mechanism is obtained. A use case scenario is presented for supporting the viability of such a mechanism.

Web services (WS) are gaining popularity for supporting business interactions in cross-organisational distributed business processes. However, current WS specifications mostly concentrate on syntactic aspects. Because multiparty collaborations in business involve complex and long-lived interactions between autonomous partners, their behaviour must be specified to ensure the reliability of the collaboration. This paper presents an event-based framework associated with a semantic definition of the commitments expressed in the event calculus, to model and monitor multi-party contracts. This framework permits to coordinate and regulate Web services in business collaborations, by allowing detection of actual and imminent violations.

Asynchronous interactions are becoming more and more important in the realization of complex B2B Web applications, and Web services are at the moment the most innovative and well-established implementation platform for communication between applications. This paper studies the existing business protocols for Web services interactions, compares their expressive power, extracts a set of patterns for implementing asynchrony, studies the trade-offs and the typical usage scenarios of the various patterns, and finally proposes a sample application that has been implemented based on these patterns. The application has been designed using a high-level modeling language for Web applications, thus showing that the studied patterns can be applied at a conceptual level as well as directly at implementation level.

E-Negotiation is the process of conducting negotiations between business partners using electronic means. The interest in e-negotiation is motivated by its potential to provide business partners with more efficient processes, enabling them to draft better contracts in less time. Most of today’s e-marketplaces support some form of e-negotiation. Numerous attempts are being made to design e-marketplaces that support more than one negotiation protocol. The main problem in designing these e-marketplaces is the lack of a systematic approach. In our view, the e-marketplace enforces negotiation protocols and therefore should make them available for consultation by humans and for automation by software agents. Separating the protocols from the e-negotiation media is a step towards a configurable e-marketplace. In this paper we address the requirements for modeling e-negotiation protocols. Then we adopt the Statechart formalism as a modeling language and provide descriptions of five commonly used e-negotiation protocols. Finally, we discuss how we move from these Statechart descriptions of the protocols to modeling the interactions between the e-marketplace participants using a web service orchestration language.

As the consumer e-commerce market grows intensively competitive, the capability of a website to capture consumers and to be accepted has been recognized as a critical issue. The user acceptance of a website not only brings immediate business opportunities, it also casts great impact on future return and loyalty buildup of the consumer. This paper is intended to explore the measurement of consumer acceptance of e-commerce website. By synthesizing previous research into a coherent body of knowledge and by recognizing the roles of contingency factors, we develop a new e-commerce website acceptance model that examines the website success. The model is extended from Garrity & Sanders Model and is expected to shed light on website design practice.

The ever-increasing popularity of the Internet has led to an explosive growth of the sheer volume of information. Recommender system is one of the possible solutions to the information overload problem. Traditional item-based collaborative filtering algorithms can provide quick and accurate recommendations by building a model offline. However, they may not be able to provide truly personalized information. For providing efficient and effective recommendations while maintaining a certain degree of personalization, in this paper, we propose a hybrid model-based recommender system which first partitions the user set based on user ratings and then performs item-based collaborative algorithms on the partitions to compute a list of recommendations. We have applied our system to the well known movielens dataset. Three measures (precision, recall and F1-measure) are used to evaluate the performance of the system. The experimental results show that our system is better than traditional collaborative recommender systems.

Evaluation of ranked queries on large text collections can be costly in terms of processing time and memory space. Dynamic pruning techniques allow both costs to be reduced, at the potential risk of decreased retrieval effectiveness. In this paper we describe an improved query pruning mechanism that offers a more resilient tradeoff between query evaluation costs and retrieval effectiveness than do previous pruning approaches.

This paper addresses the problem of semi-supervised classification on document collections using retraining (also called self-training). A possible application is focused Web crawling which may start with very few, manually selected, training documents but can be enhanced by automatically adding initially unlabeled, positively classified Web pages for retraining. Such an approach is by itself not robust and faces tuning problems regarding parameters like the number of selected documents, the number of retraining iterations, and the ratio of positive and negative classified samples used for retraining. The paper develops methods for automatically tuning these parameters, based on predicting the leave-one-out error for a re-trained classifier and avoiding that the classifier is diluted by selecting too many or weak documents for retraining. Our experiments with three different datasets confirm the practical viability of the approach.

This paper studies automatic extraction of structured data from Web pages. Each of such pages may contain several groups of structured data records. Existing automatic methods still have several limitations. In this paper, we propose a more effective method for the task. Given a page, our method first builds a tag tree based on visual information. It then performs a post-order traversal of the tree and matches subtrees in the process using a tree edit distance method and visual cues. After the process ends, data records are found and data items in them are aligned and extracted. The method can extract data from both flat and nested data records. Experimental evaluation shows that the method performs the extraction task accurately.

The prevalence of weblogs (blogs) has enabled people to share the personal experiences of tourists at specific locations and times. Such information was traditionally unavailable, except indirectly through local newspapers and periodicals. This paper describes a method of spatially and temporally obtaining specific experiences by extracting association rules from the content of blog articles. For example, we can read about visitors’ activities and evaluations of sightseeing spots. By geographically mapping their experiences, the proposed system enables observation of tourist activities and impressions of specific locations, which can often be more diverse than local guidebooks and more trustworthy than advertisements.