期刊名称：International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
印刷版ISSN：2278-1323
出版年度：2017
卷号：6
期号：12
页码：1801-1805
出版社：Shri Pannalal Research Institute of Technolgy
摘要：Opinion mining is a part of Natural Language Processing (NLP) in machine learning to tract the feelings of people’s about a single product. Opinion Mining is said to be sentiment analysis. Now a day, customers/people concentrate more on reviews to choose a product from online. With the reviews, the Customer’s decide to buy the product or not. In Partially Supervised Word Alignment Model(PSWAM), the Hill climbing EM based algorithm is used to find targets and words of opinion. The proposed method, a hierarchical Pachinko Allocation Model 2(hPAM2) is used to explore the topic relations. The Steepest Ascent Hill Climbing Algorithm is used to find the results of best from the reviews of a particular product. The steepest ascent hill climbing algorithm is used to find on the finest results of global maximum of highest degree words, decrease the error creation, its tries all possible results of the current way rather than only one results and the results are extracted out. If they cannot find on the finest results in the current way, there is a possible way to backtrack the results.
关键词：Opinion analysis; Opinion target and Opinion word extraction; Topic Model.
