A new word sense similarity measure in wordnet

摘要

Recognizing similarities between words is a basic element of computational linguistics and artificial intelligence applications. This paper presents a new approach for measuring semantic similarity between words via concepts. Our proposed measure is a hybrid system based on using a new Information content metric and edge counting-based tuning function. In proposed system, hierarchical structure is used to present information content instead of text corpus and our result will be improved by edge counting-based tuning function. The result of the system is evaluated against human similarity ratings demonstration and shows significant improvement in compare with traditional similarity measures.