Using Automated Lexical Resources In Arabic Sentence Subjectivity

Abstract

A common point in almost any work on Sentiment analysis is the need to identify which elements of language (words) contribute to express the subjectivity in text. Collecting of these elements (sentiment words) regardless the context with their polarities (positive/negative) is called sentiment lexical resources or subjective lexicon. In this paper, we investigate the method for generating Sentiment Arabic lexical Semantic Database by using lexicon based approach. Also, we study the prior polarity effects of each word using our Sentiment Arabic Lexical Semantic Database on the sentence-level subjectivity and multiple machine learning algorithms. The experiments were conducted on MPQA corpus containing subjective and objective sentences of Arabic language, and we were able to achieve 76.1 % classification accuracy.