A hybrid recommendation system and method based on user intimacy

Abstract

The invention discloses a user closeness-based mixed recommending system and a user closeness-based mixed recommending method, which are functional in recommending interesting projects for users in social network sites. The system comprises a user closeness determining module, a user closeness-based recommendation result generating module, a collaborative filtering-based recommendation result generating module, a content-based recommendation result generating module, and a result integrating module. The method is carried out by steps of determining the user closeness, acquiring user closeness-based recommendation result, acquiring collaborative filtering-based recommendation result, acquiring content-based recommendation result, and integrating the results. According to the method, data in the social network sites are fully utilized to make up the disadvantage of a traditional recommending system, therefore, the system application has the advantages of strong practicality, high accuracy and convenient implementation.

Description

technical field [0001] The invention relates to a mixed recommendation system and method based on user intimacy, belonging to the field of network data mining in the computer field. Background technique [0002] Social Network Sites (Social Network Sites, or SNS) is an online website that is convenient for people to socialize. SNS users can post photos, statuses, logs, etc. online, and other users can comment on and forward them. and other operations to enhance communication and communication between people, so as to achieve the purpose of social interaction. The current SNS generally has a recommendation function, and the purpose of the recommendation function is to recommend items that users are interested in and may accept. On the one hand, this improves the user experience and enables users to find items they are interested in faster. On the other hand, it can achieve the purpose of SNS marketing from a commercial point of view. Therefore, it is crucial to choose an ac...

AI Technical Summary

Problems solved by technology

However, due to the sparse and scattered characteristics of SNS data, the effect of simply using collaborative filtering-based recommendation or content-based recommendation is poor.

Images