RANDOM SAMPLING DATA PREPROCESSED MUTUAL INFORMATION ROCCHIO SENTIMENT CLASSIFICATION BASED SENTIMENT ANALYSIS FOR RECOMMENDATION SYSTEM

DOI: https://doie.org/10.0905/Jbse.2024105685

Dr. P.Suresh , G.Shumalatha

Keywords:

Customer reviews, Mutual Information Rocchio Sentiment Classification, Recommendation, Sentimental analysis, Term based random sampling

Abstract:

Sentiment Analysis is considered as one of significant problem to be resolved as itdetermines the opinions and emotions of the users through written contents. A Random Sampling Data Preprocessed Mutual Information Rocchio Sentiment Classification (RSDPMIRSC) Technique is proposed for providing efficient recommendation to an item with higher accuracy and lesser time. Initially, customer review and feedback of an item are collected from the large database. After that, the collected customer reviews are preprocessed in RSDPMIRSC Technique using term based random sampling process for automatically detecting and removing the stop word from the reviews. The designed process functioned through iterating the selected customer review at random manner. After removing the stop words from reviews, classification process is carried out in RSDPMIRSC Technique using Mutual Information Rocchio Sentiment Classification (MIRSC) algorithm to perform the customer review classification as positive and negative with higher accuracy and lesser time. MIRSCis a supervised method used to compute mutual information between given term and classes (i.e., positive and negative). The designed models provided the information how much the term related to the particular class. Based on the customer review classification, the recommendation regarding particular item purchase is given to the user. Experimental evaluation of RSDPMIRSC Technique is carried out on factors such as time complexity, classification accuracy, error rate and preprocessing time with respect to number of customer reviews.

PDF