Semantic Similarity Analysis on Knowledge Based and Prediction Based Models

Abstract

The similarity between two synsets or concepts is a numeral measure of the degree to which the two objects are alike or not and the similarity measures say the degree of closeness between two synsets or concepts. The similarity or dissimilarity represented by the term proximity. Proximity measures are defined to have values in the interval [0, 1]. Term Similarity, Sentence similarity and Document similarity are the areas of text similarity. Term similarity measures used to measure the similarity between individual tokens and words, Sentence similarity is the similarity between two or more sentences and Document similarity used to measure the similarity between two or more corpora. This paper is the study between Knowledge based, Distribution based and prediction based semantic models and shows how knowledge based methods capturing information and prediction based methods preserving semantic information.