A Review of Domain Adaptation for Opinion Detection and Sentiment Classification Framework

Buddhika Hasantha Kasthuriarachchy, Kasun de Zoysa, Lalith Premaratne

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 analyzing those opinions. Even though the supervised classification methods outperform human produced baselines, using it for such a framework is impractical. Since those data spans so many different domains, domain adaptation is a key feature that requires for a useful framework. Thus, 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 by domain adaptation.

Subscribe to ICTer News