A Fuzzy Logic Model for Evaluating Customer Loyalty in e-Commerce

This research proposes a model for customer's loyalty by sentiment analysis of ecommerce products. The purpose behind this research is to evaluate the response of customers in the shortest time. It practices sentiment analysis that tends to understand the user's feedback about the product and services on ecommerce sites. The data is openly available on these sites in the form of reviews, comments and appraisals. This data focus on the customer's opinions and helps business to take proficient decisions in the limited time. It takes subjective reviews because objective part contains emotion symbols. Many people do not know the proper use of emotions. It also prefers to use Stanford POS (Parts-of-Speech) tagger from Stanford Core NLP toolkit. This tagger assigns part of speech to every word of the reviews as we extract adjectives to measure the scores. This paper also used these techniques: tokenization, Lemmatization and stop words removal. By the use of soft computing approach-Fuzzy logic, it will able to design a customer loyalty model by its membership functions and truth values between 0 and 1. It uses SentiWordNet software to measure the P-N polarity scores. This proposed model reduces the problems from the related past researches. This research collects results from the reviews from Amazon. com which shows 72% customers are loyal towards ecommerce products. The outcomes that can allow business organization improve customer loyalty techniques to gain profitable results.