REVIEW CLASSIFICATION SYSTEM USING SJASM

Opinion mining has gained increasing credit and shown infinite practical worth in different years. Extracting opinion words and targets is a dominating difficulty in opinion mining. Scanning of these reviews causes computational burden for taking decision. Hence, to rule reviews automatically and summarizing them in suitable comprise is greater efficient. The distinguished suspension of producing opinion subject addresses is how to verify the sentiment, and opinion expressed in the review by all of tolerate to a numerical piece value. This paper presents a novel technique with a hybrid algorithm which combines Expectation Maximization (EM) algorithm. It meet on the main difficulty of counteraction mining called as feeling summarization. The extraction of product achievement, modern highlight value and opinion are prompt for opinion summarization as they request the performance significantly. The proposed approach consists of a software program in which mining of product achievement, technical achievement value and opinion is calculated. The main motto about software program is to extract the technical achievement value depending on revaluate, which the reviews are summarized. This software is successful for humans to get the technical values expressed in the reviews. It characterizes relations between words and targets, which is occupied to analyse the belief of each aspirant from opinion words and targets datasets. The words or targets with valuable expectation are limited in their respective datasets and the waste are moved as false results which are used to extract extraction rules. SJASM is used for combining the features and opinion words.