Implementation of Co-Occurring Phrase based Text Mining Technique for Analyzing Review Result of Product

Abstract

Customer satisfaction is important in this competitive environment manufacturer need to regularly upgrade their product and for they need to fulfil the customers need. Review is very good method from where we can identify customer’s view about product and what kind of problem they facing in using particular product. If we will not take care of this then customer will shifted towards other alternatives available in the market. But along with internet usage number of reviews is increasing every day and for that manually analysis is not effective in both the context time and cost. So it is need of some method which automatically does this work and provides us with required output. So for solving this problem we select this thesis title which will helpful to society. This thesis focus on co-occurring phrase based technique to analysis customer review of product in which we use method of keyword based filtering in first phase so that noisy reviews are filter out and save further computation. Then in second phase we propose phrase based architecture which will applied on output of first phase and further filter reviews based on negative sets of keywords. Finally we use supervised learning approach to update the sets of keywords so that we will improve system continuously.