An Hybrid Approach in Classification of Telugu Sentences

Abstract

Natural Language Processing (NLP) is a multidisciplinary research area that explores how computer understand human language in the form of text or speech to do useful things. Telugu is most prominent and morphologically rich dravidian language spoken by around one million people. When research in computational linguistics is concerns telugu is far behind other south Indian languages .Recognizing sentence similarity is most useful task in all the languages which are useful for improving plagiarism detection of documents, word sense disambiguation, query evaluation, paraphrase detection and question answering. In this paper, we discuss an hybrid approach to calculate semantic similarity score between two telugu sentences using supervised learning for classifying simple telugu verb less sentences using linguistic knowledge of telugu language and combination of rule based and stochastic methods are used to measure similarity between sentences.