2019 2018 2017 2016 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 1992 1991 1990 1989 1988 1987 1985 1984 1982 1980 1979 1978 1977 1976 1975
Journals Chapters Reports
Berger, F.C. and Bommel, P. van and Weide, Th.P. van der, Ranking strategies for navigation-based query formulation. Journal of intelligent information systems, Nr: 1, Vol: 12, 1999
Navigating through a hypermedia retrieval system bears the problem of selecting an item from a large number of options which are available to continue the trajectory. Ranking these options according to some criterion is a method to ease the task of navigation. A number of ranking strategies have already been proposed. This paper presents a formalization of the concept of ranking, and of the aforementioned strategies. Furthermore we propose two strategies which allow a personalized approach to ranking.
[ cite ]
Wondergem, B.C.M. and Bommel, P. van and Weide, Th.P. van der, Information Retrieval, een uitdagend onderzoeksgebied. Informatie Professional, Nr: 10, Vol: 3, Pages: 20-25, October, 1999, In Dutch
Recente wetenschappelijke ontwikkelingen in information retrieval (IR) worden in dit artikel door Wondergem, Van Bommel en Van der Weide aan de orde ge- steld. De drie belangrijkste modellen voor IR worden besproken en vergeleken, alsmede het doel en de gebruikte technieken van een aantal IR-projecten in Nederland. Met tot besluit: waarom een perfect IR-systeem onmogelijk is.
[ cite ]
Journals Chapters Reports
Weide, Th.P. van der, IR bijdrage aan Polyautomatiseringsboekje. PBNA, 1999, In Dutch
[ cite ]
Journals Chapters Reports
Arampatzis, A.T. and Weide, Th.P. van der and Koster, C.H.A. and Bommel, P. van, Text Filtering using Linguistically-motivated Indexing Terms. Technical report: CSI-R9901, January, Computing Science Institute, University of Nijmegen, Nijmegen, The Netherlands, EU, 1999
In this article we describe a number of text filtering experiments based on indexing terms other than simple keywords. These experiments were conducted as a first step of validating a phrase-based retrieval model. Our approach in the selection of indexing terms was based on a part-of-speech tagger and shallow parsing. Different types of indexing terms were evaluated, including keywords, nouns, verbs, adverbs, adjectives, adjacent word-pairs and head-modifier pairs. Furthermore, the effect of lemmatizing terms was investigated. Evaluation was done with three utility-based measures and average set precision. We managed to obtain experimental evidence that the phrase-based retrieval model performs better than keyword-based models. Our experiments showed that an indexing set of keywords can be reduced by retaining only some certain part-of-speech categories without any substantial negative impact in performance; in some cases the effectiveness even improved. Furthermore, we found that lemmatization can produce confusion between nouns and verbs decreasing effectiveness, while overall it is beneficial especially for phrasal terms.
[ cite ]
Jones, P.A. and Bommel, P. van and Koster, C.H.A. and Weide, Th.P. van der, Best First Search techniques in Document Processing applications. Technical report: CSI-R9902, January, Radboud University Nijmegen, 1999
[ cite ]
Wondergem, B.C.M. and Bommel, P. van and Weide, Th.P. van der, Effective and Efficient Descriptors for Information Discovery. Technical report, University of Nijmegen, 1999
The effectiveness and efficiency of searches for relevant documents strongly depend on key features of the descriptor language supported by the retrieval system. Effectiveness, for instance, is limited by the expressiveness of the descriptors. In addition, system efficiency is proportional to tractability. Furthermore, user effort is relative to comprehensibility and compactness. From a formulation point of view, compactness allows a high degree of integration, combining several occurrences of an analogous concept in a single expression. A proper descriptor language for information discovery should thus find a workable balance between these features.
Boolean index expressions (BIEs) are proposed as balanced descriptors for the communication of information needs. Compared to many other descriptor languages, an advantage of BIEs is their compactness, offering the opportunity to convey much information in a succinct way. The goal of this article is to investigate compactness of BIEs. For instance, we analyse how compactness is brought about in BIEs and how it can be effectively measured. In addition, two classes of BIEs are investigated that illustrate the bounds on the compactness of BIEs: minimal and maximal BIEs. The bounds are explicitly derived and illustrated by generic examples. Furthermore, we report on experiments that ascertain the merit of minimal and maximal BIEs.
[ cite ]
Wondergem, B.C.M. and Bommel, P. van and Weide, Th.P. van der, Counting and Generating (Boolean) Index Expressions. Technical report, University of Nijmegen, 1999, Technical Report CSI-R99xx
[ cite ]
Wondergem, B.C.M. and Bommel, P. van and Weide, Th.P. van der, Compactness of Boolean Index Expressions. Technical report: CSI-R9911, June, Radboud University Nijmegen, 1999
Boolean index expressions (BIEs) form an expressive and tractable descriptor language for text representation. Another advantage is that BIEs are compact, meaning that equivalent Boolean combinations of (regular) index epressions require more terms, connectors, and logical operators. In this article, it is investigated how compact BIEs are. Two classes od BIEs are defined that illustrate the bounds on the compactness of BIEs: minimal and maximal BIEs. The bounds are explicitly derived and illustrated by generic examples.
[ cite ]