A Survey Paper on Identifying Candidate Features in Opinion Mining using Intrinsic and Extrinsic Domain Relevance
Janhavi Suchet Vakil
Janhavi Vakil, Narsee Monjee Institute of Management Studies, Mumbai (Maharashtra), India.
Manuscript received on June 02, 2017. | Revised Manuscript received on June 05, 2017. | Manuscript published on June 15, 2017. | PP: 11-15 | Volume-4, Issue-11, June 2017. | Retrieval Number: K10390641117/2017©BEIESP
Open Access | Ethics and Policies | Cite
© The Authors. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Opinion feature extraction is the process of obtaining candidate features from the existing set of features identified from reviews and opinions. We study few techniques and propose a novel method to identify candidate features using different pattern mining approaches and extract relevant information using a set of syntactic rules. Using Dependency Parsing (DP) we can extract Parts of Speech (POS). The POS can be used to extract candidate features using syntactic rules and thus obtain candidate features. According to previous studies candidate features that are less generic and more domain-specific are then confirmed as opinion features. Previous experimental results on two real evaluation domains show that this approach may surpass several other well-established methods for identifying opinion characteristics.
Keywords: Opinion mining. Sentiment analysis, Intrinsic , Extrinsic, Domain Relevance, Stanford NLP.