Using Concept Lattice for Personalized Recommendation System Design

Caifeng Zou, Daqiang Zhang, Jiafu Wan, Mohammad Mehedi Hassan, Jaime Lloret

Abstract: A novel personalized recommendation system (PRS) based on concept lattice is proposed and used to discover valuable information according to users' requirements and interests quickly and efficiently. The system is divided into the offline part and the online part. In the offline part, the formal context and the concept lattice are constructed from the transaction database, and the association rules based on concept lattice are extracted and stored in the rule library. The new added data are used to update the concept lattice and the rule library regularly. In the online part, the behavior data of target user, the concept lattice and the rule library are used to calculate the ordered recommendation results, which are returned to the user. There are two recommendation methods in the online part, which are recommendations based on association rules and collaborative filtering recommendation. Because of the natural advantages of the concept lattice in data processing and analysis, the PRS we designed possesses better precision and faster response capability, as compared with conventional recommendation system.

External IDs:dblp:journals/sj/ZouZWHL17