Secure recommendation system for E-commerce website

Abstract

The recommendation can be done based on user interest, interpersonal influence, and interpersonal interest similarity. The social factors involved helps to recommend the product to the user in the more personalized way. The social circle includes users having similar interest to the user. E-commerce sites are providing product recommendation to the users to improve the sale of products. The recommendation is an information filtering process. Due to the accumulation of a large amount of data in a huge rate the extraction of the essential data needed is being done in data mining. Attacks on recommendation systems can be push attack or nuke attacks. The attacks can give rise to the fake recommendation which can affect the customer satisfaction. The classification of fake profiles from genuine profiles helps to improve the efficiency in the recommendation of products. This avoids the manipulation in the recommendation of products in an e-commerce website. In cold start situation where the data available for recommendation is not enough, the location attribute is included to make the recommendation.