A Hybrid Recommendation Algorithm Combing Naive Bayes Classifier and the Users’ Trust Relationship
Jianfei Li, Xiao Juan Guo, Lin Wu · 2020
With the rapid development of information technology, a huge amount of information is shown to the clients, thus leading to the so-called ‘information explosion’ or ‘information overload’. Recommend technique can follow the demand change of clients and automatically adjust the mode and contents of information service. Based on the problems of collaborative filtering recommendation of internal storage including the overdue calculation amount in large-scale system, difficult realization of real time recommendation, poor system expansibility, lack of novelty purely based on contents or the existing interests of clients, this paper proposes hybrid recommendation technique based on naive Bayesian model and client trust. On one hand, it establishes the hobby model of clients and apply Bayesian classification algorithm according to the interests of clients to recommend the classified products to clients. On the other hand, according to trust relation of clients, recommend products of higher praise to trusted clients and tap the potential interests of clients, which not only can guarantee the accuracy of recommendation, but also is of expansibility.