A Character Recognition Method Based on Particle Swarm Random Walk

Abstract

The invention relates to a character recognition method based on particle swarm random walk; firstly, the collected social network data is preprocessed; secondly, the network structure analysis is performed on the social network data, and the dynamic network subgraph snapshot based on time series is used to Structural representation of temporal network structures. Then basic features and iterative features are extracted from the high-dimensional features of nodes. The particle swarm initialization is performed on the nodes, the fitness value matrix is ​​formed from the extraction results of the temporal network structure features, and the core particles are determined according to the fitness function. Each particle is guided by the transition probability and the fitness objective function to generate different Role. Finally, the particle swarm similarity under adjacent time slices is calculated, the role distribution of adjacent time slices is predicted, and the dynamic evolution law of social networks is analyzed. The invention improves the adaptability and accuracy of the large-scale dynamic social network research method, and effectively predicts the dynamic social network evolution law and role distribution results.

Description

technical field [0001] The invention relates to a social network analysis technology, in particular to a character recognition method based on particle swarm random walk. Background technique [0002] The development of social behavior networking has accelerated the complexity and dynamics of network structures, such as information interaction networks, scientists' cooperation networks, social networks, and biological networks. The social network analysis method is a sociological analysis method that quantifies the relationship between social network actors. In social network research, role recognition is a very important research problem. It is very important for analyzing and understanding social network structure, understanding the influence of temporal evolution of network structure on roles, predicting user behavior, and studying the relationship between users and the process of information interaction. is of great significance. Behavior individuals play specific role...

AI Technical Summary

Problems solved by technology

According to the Markov logic network, Zhang et al. proposed a network opinion leader role identification method based on relational data, but lacked the consideration of time factors (Identifying network public opinion leaders based on markovlogic networks, 2014)

Character identification based on content analysis usually only focuses on personal information and network information, while ignoring the mining of potential roles. When the theme drifts, it is difficult to achieve the desired effect

Images