Research And Implementation Of Improved Label Propagation Algorithm

There are a variety of networks in real life,such as interpersonal networks,trading networks,transport networks and so on.It is of great significance to carry out community discovery in these networks,such as in the interpersonal network,you can find out the social groups with different interests and backgrounds,which is convenient to carry out different propaganda strategies;In the trading network,different communities represent different purchasing power of the customer groups,you can recommend suitable products for them;In the capital network,the community may be a potential money laundering gangs,brush drill alliance,to facilitate security services to deal with;In a similar shop network,the community can detect business,price alliance,etc.,to guide the business.In general,finding community structure in social networks can have important application scenarios.This paper mainly studies how to improve the stability and accuracy of label propagation algorithm.The two community structure is used as the starting point of the label propagation algorithm,which can limit the spread of labels between communities and improve the accuracy and stability of the label algorithm.At the same time,in order to make full use of prior knowledge,this paper proposes a novel method based on kernel community of the network,which makes full use of prior knowledge,and makes the structure of the community more obvious.In this paper,we propose a label propagation algorithm based on propagation limitation,which combine propagation limitation strategy and the use of priori.The propagation method based on propagation limit is used to validate the data set of artificial neural network,Football and Zachary.The experimental results show that,compared with the original label propagation algorithm,the use of the two community structure can effectively limit the spread of labels in the community.The two methods have the same performance for the obvious structure.When the community structure is not obvious,the label propagation algorithm based on the propagation constraints can effectively discover the community structure.Therefore,the label propagation algorithm based on propagation constraints has improved the instability of the original label propagation algorithm and the problem of community aggregation.