Calculating the Similarity of Short Ver...

Over the last decade, Similarity Calculation (SC) has become a topic of major interest in the realm of Natural Language Processing (NLP), including practical domains like Question Answering (QA) and language understanding, which is of interest amongst others for controling intelligent robots. In the field of Word Sense Disambiguation (WSD), a series of different algorithms have been developed for measuring similarity, Lesk being certainly among the most prominent. As opposed to many applications featuring similarity calculation of nouns, in this paper we rise the question whether the calculation of verbs and short verb phrases can be done in a similar way. The result is, that measuring the similarity of verbs and short verbal phrases has to be tackled employing a completely different strategy. In fact, best results can be obtained with Lesk as well, provided that filtering the input of rules and the number of processing steps are well differentiated with respect to the rather simple calculation of noun similarities.