A review of domain adaptation for opinion detection and sentiment classification

Social networks and micro-blogging sites have become a treasured source to reveal “What other people think”. However, the usage is quite limited without sophisticated framework for mining and analysing those opinions. Even though the supervised classification methods outperform human produced baselines, using it for such a framework is impractical, mainly due to the fact that those data spans so many different domains. Thus, domain adaptation is a key feature that requires for a useful framework. Hence, this paper discusses existing works on domain adaptation in the context of opinion detection and sentiment classification. It focuses the areas covered by reviewed frameworks and evaluates papers, based on important parameters. Set of experiments are also performed to compare the effect on classification accuracy due to domain adaptation.