Abstract: Data mining - an analytical process designed to explore data in which the opinion mining deals with the computational treatment of opinion, sentiment and subjective in text. The main application of opinion mining is collecting the online reviews about the product, social networks informal text. The research problem is extracting the opinion targets and the opinion words and detecting the opinion relations among the words. A novel approach based on the partially supervised alignment model for identifying the opinion relations as an alignment process have been proposed to satisfy the long span relations. To precisely mine the opinion relations among words, the Word Alignment Model (WAM) is used and to progress the error propagation, the graph based co-ranking algorithm is motivated. By comparing with the syntax based method, the word alignment model effectively reduces the parsing errors and the co-ranking algorithm decreases the error probability. The datasets CRD, COAE 2008 and Large are used in various methods. The survey shows the algorithm effectively outperforms when compare to previous methods.

Keywords: Data mining, Opinion mining, WAM, Opinion word, Opinion target.