Online product prediction and recommendation using probability graphical model and collaborative filtering: A new approach

Prediction systems apply knowledge discovery techniques to the problem of making personalized product recommendations. The tremendous growth of customers and products in recent years, poses some key challenges for prediction systems, as these are producing high quality recommendations per seconds for millions of customers and products. New recommender system technologies are needed that can quickly produce quality recommendations, even for very large-scale problems. One of the most successful recommender technologies to date is automatic collaborative filtering (CF). Collaborating systems works by measuring distances between people in “taste space”, and predicting interest in untried items based on a weighted sum of nearby users impressions of the untried items. This paper presents a new and efficient approach that works using Bayesian belief networks (BBN) and that calculate the probabilities of inter-dependent events by giving each parent event a weighting (Expert systems). Nearest-neighbor collaborative filtering provides a successful means of generating recommendations for web users. Finally, we explore the ability of our method to generate useful recommendations, then reporting the results of a user-study, where users prefer the recommendations generated by our approach.