Identification of Assamese Question sentence using SVM and Naïve Bayes

Abstract

This paper presents an approach to identify Assamese question sentence from an corpus. It is an important step towards question sentence detection from a document. In this paper, the question structure of different question type in Assamese language is analyzed. We deliberated the lexical feature, syntactic feature and semantic features of question. Syntactic and lexical feature is used to identify Assamese question. We have used SVM, Naïve Bayes for categorization Assamese sentence using syntactic and lexical pattern. We have created our own dataset for training and testing and get maximum of 92% accuracy.