Toward Better Recommender System by Collaborative Computation with Privacy Preserved

Recommender systems are best known for the usage on E-commerce websites, with the aim of helping customers in the decision making and product selection process by providing a list of recommended items. Since most of the recommender systems of E-commerce websites suffer from data scarcity, joining recommender system databases is said to improve the prediction and recommendation results. However, there will be a risk of revealing the raw customer-product information by sharing the databases between websites. In this research, a new scenario of collaborative computation is proposed. A very preliminary result has shown the advantage of joining recommender systems. In addition, private user raw data is preserved while doing collaborative computation for better recommendations.