Learning a domain-independent classifier for effective sentiment classification: A gloss-based approach

With the success and explosion of Web 2.0 applications, consumers are not only shopping and comparing products on the Web but also providing their product reviews on the Internet platform. Such consumer reviews are essential and beneficial for customers, merchants, and product manufacturers. However, as the number of consumer reviews expands rapidly, it becomes more difficult for users to obtain a comprehensive view of consumer opinions pertaining to the products of interest through a manual analysis. The development of techniques capable of automatic analysis and summary of the sentiments of consumer reviews pertaining to specific products becomes desirable and essential. In this study, we concentrate on the sentiment classification of consumer reviews. To address the problem of domain dependency commonly encountered by existing techniques, we proposed a gloss-based sentiment classification technique which is domain-independent in nature. According to our empirical evaluation results, the proposed gloss-based sentiment classification technique outperforms the traditional approach.