Taxonomy Learning - Factoring the Structure of a Taxonomy into a Semantic Classification Decision

Taxonomy learning – factoring the structure of a taxonomy into a semantic classification decision Viktor PEKAR Bashkir State University Ufa, Russia, 450000 vpekar@ufanet.ru Steffen STAAB Institute AIFB, University of Karlsruhe http://www.aifb.uni-karlsruhe.de/WBS & Learning Lab Lower Saxony http://www.learninglab.de Abstract The paper examines different possibilities to take advantage of the taxonomic or- ganization of a thesaurus to improve the accuracy of classifying new words into its classes. The results of the study demon- strate that taxonomic similarity between nearest neighbors, in addition to their dis- tributional similarity to the new word, may be useful evidence on which classifi- cation decision can be based. 1. Introduction Machine-readable thesauri are now an indispen- sable part for a wide range of NLP applications such as information extraction or semantics- sensitive information retrieval. Since their man- ual construction is very expensive, a lot of recent NLP research has been aiming to develop ways to automatically acquire lexical knowledge from corpus data. In this paper we address the problem of large- scale augmenting a thesaurus with new lexical items. The specifics of the task are a big number of classes into which new words need to be clas- sified and hence a lot of poorly predictable se- mantic distinctions that have to be taken into account. For this reason, knowledge-poor ap- proaches such as the distributional approach are particularly suited for this task. Its previous ap- plications (e.g., Grefenstette 1993, Hearst and Schuetze 1993, Takunaga et al 1997, Lin 1998, Caraballo 1999) demonstrated that cooccurrence statistics on a target word is often sufficient for its automatical classification into one of numer- ous classes such as synsets of WordNet. Distributional techniques, however, are poorly applicable to rare words, i.e., those words for which a corpus does not contain enough cooc- currence data to judge about their meaning. Such words are the primary concern of many practical NLP applications: as a rule, they are semanti- cally focused words and carry a lot of important information. If one has to do with a specific domain of lexicon, sparse data is a problem par- ticularly difficult to overcome. The major challenge for the application of the distributional approach in this area is, therefore, the development of ways to minimize the amount of corpus data required to successfully carry out a task. In this study we focus on opti- mization possibilities of an important phase in the process of automatically augmenting a the- saurus – the classification algorithm. The main hypothesis we test here is that the accuracy of semantic classification may be improved by taking advantage of information about taxo- nomic relations between word classes contained in a thesaurus. On the example of a domain-specific thesaurus we compare the performance of three state-of- the-art classifiers which presume flat organiza- tion of thesaurus classes and two classification algorithms, which make use of taxonomic or- ganization of the thesaurus: the "tree descend- ing" and the "tree ascending" algorithms. We find that a version of the tree ascending algo- rithm, though not improving on …

Content truncated to 500 words