Multi-relation social network model mining method based on Bayesian method

Abstract

The invention provides a multi-relation social network model mining method based on the Bayesian method. The method comprises the following steps: performing uncertain relation modeling based on an n-cell fuzzy number for aggregating a plurality of interactive relationships between social networks and keeping the subjection precision of a fuzzy subset as much as possible under inaccurate and incomplete sampling environment; and then generating a zero model and an enumeration subgraph by theoretical analysis, and performing motif recognition based on the Bayesian method to accelerate the algorithm convergence process. By adoption of the method, relatively universal description of real social networks can be realized, and in particular the description of multi-relation and uncertain relation. On the basis, the existing general motif recognition method is expanded to mine the non-trivial interaction mode for providing forceful support for analysis and mining of the social network, for example, public opinion transmission, recommendation system, precision marketing and the like, on the premise of ensuring the precision and reliability.

Description

technical field [0001] The present invention relates to a multi-relationship social network pattern mining method based on Bayesian method, which is oriented to multi-relationship social network, and aims to mine the interaction pattern through the identification of network weighted motifs, so as to analyze the mesoscopic structure of the community, Clarifying the nature of communities and subdividing communities to provide more precise support belongs to the technical field of network weighted motif mining. Background technique [0002] For multi-relational social networks, existing motif mining techniques are mainly based on simple networks, that is, weightless models. [0003] Traditional multi-relational modeling often adopts the weighted average method, which is based on the ideal presupposition of complete information. In reality, due to the imperfection of data set scale and collection technology, the weighted average method cannot comprehensively and effectively cha...

AI Technical Summary

Problems solved by technology

[0007] The technical problem to be solved by the present invention is: aiming at the lack of singleness and uncertainty of the existing general motif recognition methods, to provide a more general description of the real social network, especially for multiple Relationships and uncertain relationships can provide high-precision, high-reliability description and mining methods

Images