An Experimental Technique on Features Extraction for Product Feedback using Opinion Mining
Jawahar Gawade1, Latha Parthiban2
1Jawahar Gawade, Ph.D. Scholar: Computer Science Engineering Bharath University, Chennai, India.
2Latha Parthiban, Associate Professor, Department of Computer Science, Pondicherry University, CC, India.
Manuscript received on 09 June 2019 | Revised Manuscript received on 14 June 2019 | Manuscript Published on 08 July 2019 | PP: 481-481 | Volume-8 Issue-8S3 June 2019 | Retrieval Number: H11100688S319/19©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. 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: Now a days with increased use of social media, user reviews on any product plays a significant role. The existing opinion mining techniques operate only on single review corpus; it doesn’t consider distribution of opinion features across different corpora. A new technique is proposed to extract opinion features from online reviews across two corpora, one a given review corpus and other is contrasting corpus. This discrepancy is evaluated using domain relevance. The first step is to find candidate opinion features in domain review corpus using syntactic dependence rules. This evaluate intrinsic domain relevance scores for each candidate on domain-dependent corpora and extrinsic domain relevance score on domain-independent corpora. A candidate features which are less generic and more domain specific are the final opinion features. This interval thresholding is called as the intrinsic and extrinsic domain relevance.
Keywords: Corpus, Syntactic Rules, Domain Relevance, Candidate Feature, opinion mining, opinion featur
Scope of the Article: Data Mining