Web based prediction and recommendation of products in electronic commerce using association rule learning and genetic algorithm

摘要

Recommendation systems make suggestions about artifacts or products to a user. For instance, they may predict whether a user would be interested in purchasing a particular product. Conventional social recommendation systems collect ratings of artifacts from many individuals, and use nearest-neighbor techniques to make recommendations to a user concerning new artifacts. This paper presents a new method for recommending items/products to users based on customer's opinions. The proposed method is a variation of traditional collaborative technique, where predictions and recommendations are computed using a set of customer's opinions from an independent database. The prediction and recommendation is done by using weighted opinions according to their similarity to the user and association rule learning. To achieve better result, the database is optimized by excluding the weak entities using Genetic Algorithm (GA). The final prediction and recommendation is done using the association rule learning on the optimized database. The efficiency of the proposed method is evaluated on a product database of 10 products having 4,500 customer's opinions. The experimental results show that the proposed method overcomes some of the problems associated with collaborative filtering methods, as reported in the literature.