Integrating Multiple Linear Regression and Multicriteria Collaborative Filtering for Better Recommendation
Chein-Shung Hwang, Yu-Cheng Kao, Ping Yu · 2010
Recommender systems are emergent to help overcome the information overload challenges by providing personalized suggestion based on users' preference. To achieve this goal, most recommender systems utilize Collaborative Filtering (CF) technique. Multiple Criteria Decision Analysis (MCDA) is a discipline aimed at supporting decision makers to make an optimal selection in an environment of conflicting and competing criteria. In the paper, we propose a mechanism for integrating MCDA into the CF process for multiple criteria recommendations. The proposed system consists of two main parts. Firstly, the weight of each user toward each feature is computed by using multiple linear regression. The feature weight is then incorporated into the collaborative filtering process to provide recommendations. The experimental results showed that the proposed approach outperformed the single criterion CF method.