Dihedral Product Recommendation System for E-commerce Using Data Mining Applications

With the development of communication networks, online access to information among much dense similar information has become a big problem. Therefore finding customers’ desired products in e-commerce is more difficult now. Product Recommendation System (PRS)tries to solve this problem and reduce the overhead of communication networks by giving rec-ommendations to customers. The purpose of this study is clustering products and create groups of products that have similar characteristics.Thus access to products with common attributes be-comes easier and it prevents customers from searching in confusion or wasting time. In this arti-cle, data collected from electronic stores is clustered and grouped using C-Means algorithm. An-other goal is to predict whether or not the customers purchase accessories related to the products they tend to buy. To explore the relationship between products, you must use the customer's be-havior and their purchase history using association rules .These rules use data mining to discover the relationship. This relationship eventually will lead to recommendations to customers when they purchase the product The relationship between products helps to increase the accuracy of recommendations and also increases the likelihood of selling related products in electronic trans-actions .More detailed recommendations will lead to carefully selected customers. The results indicate that the accuracy of recommendations in the proposed PRS is more than other RPSs.

کلیدواژه ها:

product recommendation system ، dihedral RPS ، e-commerce ، clustering ، associa-tion rules