UPV-SI: Word Sense Induction using Self Term Expansion

Proceedings of the 4th International Workshop on Semantic Evaluations (SemEval-2007), pages 430–433, Prague, June 2007. c⃝2007 Association for Computational Linguistics UPV-SI: Word Sense Induction using Self Term Expansion∗ David Pinto(1,2) and Paolo Rosso1 1Polytechnic University of Valencia DSIC, Valencia, Spain, 46022 2B. Autonomous University of Puebla FCC, Puebla, Mexico, 72570 {dpinto, prosso}@dsic.upv.es H´ector Jim´enez-Salazar Autonomous Metropolitan University Department of Information Technologies Cuajimalpa, DF, Mexico, 11850 hgimenezs@gmail.com Abstract In this paper we are reporting the re- sults obtained participating in the “Eval- uating Word Sense Induction and Dis- crimination Systems” task of Semeval 2007. Our totally unsupervised system performed an automatic self-term expan- sion process by mean of co-ocurrence terms and, thereafter, it executed the unsupervised KStar clustering method. Two ranking tables with different eval- uation measures were calculated by the task organizers, every table with two baselines and six runs submitted by dif- ferent teams. We were ranked third place in both ranking tables obtaining a better performance than three different baselines, and outperforming the average score. 1 Introduction Word Sense Disambiguation (WSD) is a partic- ular problem of computational linguistics which consists in determining the correct sense for a given ambiguous word. It is well-known that su- pervised algorithms have obtained the best re- sults in public evaluations, but their accuracy is close related with the amount of hand-tagged data available. The construction of that kind of training data is difficult for real applications. The unsupervised WSD overcomes this draw- back by using clustering algorithms which do ∗This work has been partially supported by the MCyT TIN2006-15265-C06-04 project, as well as by the BUAP- 701 PROMEP/103.5/05/1536 grant not need training data in order to determine the possible sense for a given ambiguous word. This paper describes a simple technique for unsupervised sense induction for ambiguous words. The approach is based on a self term ex- pansion technique which constructs a set of co- ocurrence terms and, thereafter, it uses this set to expand the target dataset. The implemented system was performed in the task “SemEval- 2007 Task 2: Evaluating Word Sense Induc- tion and Discrimination Systems”(Agirre and A., 2007). The aim of the task was to per- mit a comparison across sense-induction and dis- crimination systems. Moreover, the comparison with other supervised and knowledge-based sys- tems may be also done, since the test corpus was borrowed from the well known “English lexical- sample” task in SemEval-2007, with the usual training + test split. The self term expansion method consists in replacing terms of a document by a set of co- related terms. The goal is to improve natu- ral language processing tasks such as cluster- ing narrow-domain short texts. This process may be done by mean of different ways, of- ten just by using a knowledge database. In information retrieval, for instance, the expan- sion of query terms is a very investigated topic which has shown to improve results with respect to when query expansion is not employed (Qiu and Frei, 1993; Ruge, 1992; R.Baeza-Yates and Ribeiro-Neto, 1999; Grefenstette, …

Content truncated to 500 words