Sentence similarity detection in Malayalam language using cosine similarity

Identifying paraphrase in Malayalam language is difficult task because it is a highly agglutinative language and the linguistic structure in Malayalam language is complex compared to other languages. Here we use individual words synonyms to find the similarity between two sentences. In this paper, cosine similarity method is used to find the paraphrases in Malayalam language. In this paper we present the observations on sentence similarity between two Malayalam sentences using cosine similarity method, we used test data of 900 and 1400 sentence pairs of FIRE 2016 Malayalam corpus that used in two iterations to present and obtained an accuracy of 0.8 and 0.59.