Scoring products from reviews through application of fuzzy techniques

摘要

e-Commerce is becoming an increasingly popular trend nowadays, where a consumer can buy a product through the internet. After buying a product, consumers can post their reviews and comments about the product in web sites and blogs. These comments can be a powerful source of finding out the consumer preferences and making recommendations of products to a new consumer who desires to buy a product. Even though there are existing systems which try to utilize these reviews and try to find the inherent quality of a product, a complete and comprehensive system is absent which provides a mechanism to rate products without any bias. Hence, an attempt is made to propose a complete system which starts from reviews collection to the analysis of reviews and the final score calculation of a product. The novel methods used in this system are the calculation of spam level of each review and the calculation of scores for each feature of a product. Fuzzy logic is used to calculate the spam level scores and the product ratings.