Hybrid filtering recommendation method and system based on duality theory

Abstract

The invention relates to a hybrid filtering recommendation method and system based on a duality theory. The method comprises: obtaining an original probability set based on a recommendation list according to item similarity and preference vectors of users; based on the item similarity and the preference vector, determining an original item recommendation set and an original similar user cluster, and obtaining a dual probability set according to a user similarity matrix, an item scoring matrix and a recommendation list; based on the item scoring matrix, determining a current dual item recommendation set and a dual similar user cluster; and obtaining a final recommendation list according to the original item recommendation set, the dual item recommendation set, the original similar user cluster, the dual similar user cluster, the original probability set and the dual probability set. Under the condition that historical access data and evaluation data are extremely sparse, the recommendation performance of the hybrid recommendation method can be improved, and then the accuracy of a recommendation result is improved.

Description

technical field [0001] The present invention relates to the technical field of recommendation, in particular to a hybrid filtering recommendation method and system based on dual theory. Background technique [0002] With the explosive growth of Internet resources, especially commodities and entertainment resources, more and more people search for the items they need online, and recommender systems have become a research hotspot in e-commerce and other fields and have been widely used. [0003] Hybrid recommendation techniques combine two or more recommendation techniques to achieve better performance. Because some recommendation systems can work effectively when there is historical access data, such as content-based recommendation; while some recommendation systems can work effectively when there is enough evaluation data, such as collaborative filtering recommendation. [0004] In the current research on hybrid recommendation technology, strategies such as neural network, ...

AI Technical Summary

Problems solved by technology

The implementation of these methods is based on a premise that there must be a lot of user history data and evaluation data. When faced with the situation of extremely sparse historical access data and evaluation data, the recommendation performance of the hybrid recommendation method is low, which leads to the failure of personalized recommendation. less accurate

Images