A Framework For Sentiment Analysis On Smartphone Application Stores

Abstract

This paper presents a framework for extracting customer opinions from review comments of smartphone application stores based on sentiment analysis. We focus on constructing a sentiment dictionary, generating opinion units from the review text, and further determining the sentiment polarity of each discovered opinion unit. We extract opinion units by traversing a syntactic tree which is built from the text. Also, we take into consideration context-dependent sentiment, which indicates different polarities with respect to different features. Overall, the framework consists of four major components: lexicon recommendation module, dictionary management module, opinion unit detection module, and sentiment polarity decision module. These modules are implemented as a self-contained software package.