UPV-WSD : Combining different WSD Methods by means of Fuzzy Borda Voting

Proceedings of the 4th International Workshop on Semantic Evaluations (SemEval-2007), pages 434–437, Prague, June 2007. c⃝2007 Association for Computational Linguistics UPV-WSD : Combining different WSD Methods by means of Fuzzy Borda Voting Davide Buscaldi and Paolo Rosso DSIC, Dpto. Sistemas Inform´aticos y Computaci´on Universidad Polit´ecnica de Valencia Valencia, Spain {dbuscaldi,prosso}@dsic.upv.es Abstract This paper describes the WSD system devel- oped for our participation to the SemEval-1. It combines various methods by means of a fuzzy Borda voting. The fuzzy Borda vote- counting scheme is one of the best known methods in the field of collective decision making. In our system the different disam- biguation methods are considered as experts that give a preference ranking for the senses a word can be assigned. Then the prefer- ences are evaluated using the fuzzy Borda scheme in order to select the best sense. The methods we considered are the sense fre- quency probability calculated over SemCor, the Conceptual Density calculated over both hyperonyms and meronyms hyerarchies in WordNet, the extended Lesk by Banerjee and Pedersen, and finally a method based on WordNet domains. 1 Introduction One of the lessons learned from our previous experi- ence at Senseval-31 (Buscaldi et al., 2004; Vazquez et al., 2004) is that the integration of different sys- tems usually works better than a standalone system. In our opinion this reflects the reality where humans do not apply always the same rule in order to disam- biguate the same ambigue word; for instance, if we consider the sentences “He hit a home run” and “The thermometer hit 100 degrees”, in the first case the sport domain helps in determining the right sense for 1http://www.senseval.org hit, whereas in the latter the disambiguation is car- ried out mostly depending on the fact that the subject of the sentence is an object. The combination of distinct methods represents itself a major problem. If the methods return dif- ferent answers, how can we select the best one? In this sense the available choices are the following: • Rule-based selection: a set of rules that can be both hand-made or automatically learned from examples; • Probability-based: the output of the methods is normalized in the range [0, 1] and is considered as a probability. Then the values are multiplied in order to obtain the sense with a maximum probability. • Vote-based: the output of the methods is con- sidered as a weighted vote. Then a voting scheme is used in order to obtain the most voted sense. In our previous participation with the R2D2 project (Vazquez et al., 2004) the selection was rule-based, with hand-made rules that attempted to take into ac- count the reliability of the various method. We sub- sequently attempted to learn automatically the rules, but the results of these experiments did not allow to determine clearly which method was to be used in each context. Working with probabilities can be problematic due to the null probabilities that make necessary the adoption of smoothing techniques. Therefore, we opted for a voting scheme, …

Content truncated to 500 words