Enhanced Collaborative Filtering: A Product Life Cycle Approach


Hyun Sil Moon
Young U. Ryu
Jae Kyeong Kim


Recommender systems are ubiquitous not only among e-commerce enterprises but also among various brick-andmortar firms. Popular collaborative filtering-based recommender systems use only individual customers' preferences discovered in their profiles containing historical purchase (or similar) records. On the other hand, the market trends of products are another factor that can substantially affect the likelihood of products being adopted. Consequently, there are rooms for further improvements in collaborative filtering-based recommendation. In this study, we propose the use of the product life cycle concept based on the Bass model and suggest an approach that integrates the general popularity effect (market trend) and the individual preference effect in order to improve recommendation effectiveness of collaborative filtering. Through experimental validation, we find that our approach of combining the product life cycle concept and collaborative filtering performs better than the approach based on typical user-based collaborative filtering alone. In addition, the experiment results show that the influence of preference and popularity effects may vary based on market characteristics. Consequently, the proposed approach can be used as a marketing tool functioning as a basis for valuable services to customers.

Key Word: 

Published Date: 

August, 2019

Full File: