A social network friend recommendation method based on community discovery

Abstract

The invention discloses a social network friend recommendation method based on community discovery, belongs to the fields of data mining, social network, etc., and solves the problem that the characteristics of the social network are not considered in the existing personalized recommendation. The present invention collects user historical data in a social network, models the user's hobbies, and obtains a set of preference vectors for all users; according to the user preference vector set representing the user's hobbies and hobbies and the user's friend relationship preference vector, the user is aggregated Class, find out the user's hobbies and friend relationship overlapping community in the social network; according to the user's hobbies and user's friend relationship overlapping community, get the target user's initial list of friends to be recommended, and the obtained initial list of friends to be recommended Filter and sort to get the final friend list to be recommended. The invention utilizes the user's hobbies and friend relationships to recommend friends, and is more suitable for social networks.

Description

technical field [0001] A social network friend recommendation method based on community discovery, which uses the typical complex network characteristics of social networks to find the user's hobbies and hobbies in the social network, uses the user's hobbies and friend relationships to recommend friends, and improves the accuracy of personalized recommendation User experience of degree and social services, involving data mining, social network, personalized recommendation and other fields. Background technique [0002] Social network is a typical complex network, which has small-world, scale-free and community structure characteristics of complex networks. [0003] The small-world property means that in the case of a large-scale network, any two network nodes can establish a connection through relatively small steps. The social network not only fully conforms to the "six degrees of separation" theory, but also has a small characteristic path length, which conforms to the sm...

AI Technical Summary

Problems solved by technology

[0009] Existing community discovery algorithms usually only use the network topology graph of user-friend relationships in social networks, and the communities found are non-overlapping communities

In addition, the existing friend recommendation methods often only use the user's friend relationship, without considering the user's hobbies or the dynamic changes of the user's hobbies.

Images