Catálogo de publicaciones - libros

Current research in large-scale information management systems is focused on unsupervised methods and techniques for information processing. Such approaches support scalability in regard to present-day exponential growth in information processing needs. In this paper we focus on the problem of automated quality evaluation of a completely unsupervised metadata extraction process in the Digital Libraries domain. In particular, we investigate resulting metadata quality applying specific extraction methodology for scientific documents. We propose and discuss precise quality metrics and measure the dynamics of such quality metrics as a function of the extracted information from the repository and size of the repository.

The University of Trier maintains the DBLP (igital ibliography & ibrary roject) Computer Science Bibliography which offers bibliographic information about more than 870.000 scientific publications. This paper describes the , a meta search engine that is able to search for full text publications in PDF format for each DBLP entry on the web. Various search engines such as Google and Yahoo are used as data sources. The retrieved documents are additionally analysed and ranked according to their relevance. The proposed system differs from systems like CiteSeer in so far, that the DBLP Webcrawler builds upon metadata and tries to find relevant full-texts whereas CiteSeer mainly starts with full-texts and extracts metadata.

Service models for digital libraries have looked into how services may be decomposed into modules and components for greater flexiblity. These models are, however, mostly aimed at server-side applications. With the emergence of Ajax and similar techniques for processing XML documents within a Web browser, it has now become feasible for a browser to perform far more of the computational tasks traditionally encompassed in server-side DL services. Among other advantages, moving computation to the client can result in improved performance and scalability. As a new twist on service oriented computing, it is argued in this paper that digital library services can be provided partially or wholly through applications that execute client-side. Two case studies are provided to illustrate that such in-browser services are feasible and in fact more powerful and flexible than the traditional server-side service model.

This work describes 5SQual, a quantitative quality assessment tool for digital libraries based on the 5S framework. 5SQual aims to help administrators of digital libraries during the implementation and maintenance phases of a digital library, providing ways to verify the quality of digital objects, metadata and services. The tool has been designed in a flexible way, which allows it to be applied to many systems, as long as the necessary data is available. To facilitate the input of these data, the tool provides a wizard-like interface that guides the user through its configuration process.

Optical music recognition (OMR) enables librarians to digitise early music sources on a large scale. The cost of expert human labour to correct automatic recognition errors dominates the cost of such projects. To reduce the number of recognition errors in the OMR process, we present an innovative approach to adapt the system dynamically, taking advantage of the human editing work that is part of any digitisation project. The corrected data are used to perform MAP adaptation, a machine-learning technique used previously in speech recognition and optical character recognition (OCR). Our experiments show that this technique can reduce editing costs by more than half.

The (SOM) has been proposed as an interface for exploring Digital Libraries, in addition to conventional search and browsing. With advanced visualisations uncovering the contents and its structure, and advanced interaction modes as zooming, panning and area selection, the SOM becomes a feasible alternative to classical interfaces. However, there are still shortcomings in helping the user to understand the map – there are insufficient methods developed for describing the map to support the user in the analysis of the map contents. In this paper, we present recent work in assisting the user in exploring the map by automatically describing maps using advanced labelling and summarisation of map regions.

We present the results of a questionnaire survey for the access and reproduction policies of 67 digital collections in 34 libraries (national, academic, public, special etc) from 13 countries. We examine and analyze the above policies in relation to specific factors, such as, the acquisition method, copyright ownership, library type (national, academic, etc.), content creation (digitized, born-digital) and content type (audio, video, etc.); how these factors affect the policies of the examined digital collections. Responses were received from a range of library sectors but by far the best responses came from academic libraries, in which we focus. We extract policy (access, reproduction) rules and alternatives according to these factors that lead to a policy decision tree on digital information management for academic libraries. The resulting decision tree is based on a policy model; the model and tree are divided into two parts: for digitized and born-digital content.

Current digital libraries and online bibliographies share several properties with the Web and thus also share some of its problems. Faceted classifications and Semantic Web technologies are explored as possible approaches to improving digital libraries and alleviating their respective shortcomings. We describe the possibilities of using faceted navigation and its personalization in digital libraries. We propose a method of faceted browser adaptation based on an automatically acquired user model with support for dynamic facet generation.

This poster reports on a study carried out to investigate and analyze a specific category of digital library visual interfaces that support information seeking, exploration and retrieval based on metadata representations, namely . This study has examined 21 metadata-enhanced digital library visual interfaces from the following perspectives: a) information access and retrieval features supported; b) metadata elements used; c) visualization techniques and metaphors utilized. The results show that visual interfaces to digital libraries enhanced with metadata are becoming more widespread. The study also demonstrates that the combined use of visualization techniques and metaphors is becoming increasingly prevalent as a design strategy to support users’ information exploration.

This paper describes the development of a distributed annotation system which enables collaborative document consultation and creates new access to otherwise hard to index digital documents. It takes the annotations one step further: not only the same types of annotations are available across file formats, but robust references to the documents introduce format and location independence, and enable the attachment even when the document has been modified. These features are achieved using standards of the digital library systems, and don’t require modification of the original documents or impose further restrictions, thus being infrastructure independent. Integration into the Kepler workflow system allows annotating workflow results, and the automatic creation and indexing of annotations in document oriented workflows, which can be used as a flexible way to archive and index collections in the Cheshire3 search engine.