Automatic Extraction of Semantic Relations by Using Web Statistical Information

A semantic network is a graph which represents semantic relations between concepts, used in a lot of fields as a form of knowledge representation. This paper describes an automatic approach to identify semantic relations between concepts by using statistical information extracted from the Web. We automatically constructed an associative network starting from a lexicon. Moreover we applied these measures to the ESL semantic similarity test proving that our model is suitable for representing semantic correlations between terms obtaining an accuracy which is comparable with the state of the art.