OPINION FEATURES EXATCTION USING INTRINSIC AND EXTRINSIC DOMAIN RELEVANCE

An Sentiment analysis, also called opinion mining, is the field of study that analyzes people’s opinions, sentiments, attitudes and emotions towards entities such as products, services, organizations, and their attributes. Mining of opinions from customer reviews is received tremendous attention from both domain dependent document and domain independent document as it decides the overall rating of any product.. Opinion identification is not big problem if we use a single review corpus, but it will give poor results.. In this current paper we propose Novel technique for mining opinion features from two or more review corpus. This technique use two corpus one domain dependent and other domain independent. We will major domain dependent relevance for candidate feature with both domain dependent and domain independent corpus, we call it as intrinsic domain relevance and extrinsic domain relevance respectively. The opinion features with IDR greater than intrinsic domain relevance threshold and less than extrinsic domain relevance are user opinions plays an important role in finding grade of the product. Many users require grade of the product along with which positive and negative factors decide this rating. In proposed paper different techniques are proposed to extract opinion features from two or more review corpora. KEYWORDS: Information search and retrieval, natural language processing, opinion mining, opinion feature