Selection Preference Basede Verb Sense Disambiguation Using WordNet

Selectional Preference Based Verb Sense Disambiguation Using WordNet Patrick Ye Department of Computer Science and Software Engineering University of Melbourne, VIC 3010, Australia jingy@cs.mu.oz.au Abstract Selectional preferences are a source of linguistic infor- mation commonly applied to the task of Word Sense Dis- ambiguation (WSD). To date, WSD systems using se- lectional preferences as the main disambiguation mech- anism have achieved limited success. One possible rea- son for this limitation is the limited number of semantic roles used in the construction of selectional preferences. This study investigates whether better performance can be achieved using the current state-of-art semantic role labelling systems, and explores alternative ways of ap- plying selectional preferences for WSD. In this study, WordNet noun synonym sets and hypernym sets were used in the construction of selectional preferences; Sem- cor2.0 data was used for the training and evaluation of a support vector machine classifier and a Naive Bayes classifier. 1 Introduction Word Sense Disambiguation (WSD) is the process of examining word tokens in a given context and specifying exactly which sense of each word is in- tended in that context. It has many applications in natural language processing related areas such as document retrieval, question answering, and com- positional sentence analysis (Jurafsky and Martin, 2000), to name a few. WSD systems can be roughly divided into two categories based on how the disambiguation infor- mation is obtained and applied: knowledge based and corpus based. Knowledge based systems in general require certain existing linguistic informa- tion repositories which provides all the information that can be used by the disambiguation system to distinguish different senses of the same polysemous words based on the context. Examples of knowl- edge based systems include dictionary based sys- tems (Lesk, 1986) and selectional preference based systems (Resnik, 1997). Corpus based systems in general do not require any linguistic information, instead, they require a certain amount of training data (labelled or unla- belled), and a set of predefi ned disambiguation fea- tures which can be used by a statistical method to train a classifi er which then is used in the disam- biguation of previously unseen data. A corpus based system is described in (Yarowsky, 1995). Selectional preferences between predicating words (verbs and adjectives) and their arguments (nouns) are a type of linguistic information which has previously been combined with statistical methods to perform word sense disambiguation, ((Resnik, 1997) and (McCarthy and Carroll, 2003)). A selectional preference is a function mapping semantic-role to noun type. The basic assumption made by all selectional preference based WSD systems is that the different senses of the same predicating word would have different selectional preferences with their arguments. As will be discussed in section 2.2, selectional preference based WSD systems developed so far are limited in terms of coverage and accuracy. In my opinion, the most important cause of this lim- itation is these systems’ inability to extract a suffi - cient number of semantic roles to be used in the con- struction of selectional preferences. For example, if …

Content truncated to 500 words