Research On The Method Of Constructing The Character Map Based On Micro-blog

With the rapid development of the Internet,more and more users are involved in the Internet.A large amount of data is generated every day on the Internet,which contains a lot of useful information.How to extract structured data from unstructured text is the focus of this paper.However,in natural language documents,a large number of character social relationships are described.It is very useful to extract characters' social relations from these documents for analyzing the characters' social relations.The bootstrapping relation extraction system can be applied to the micro-blog environment effectively.Based on the bootstrapping relation extraction system,four improvements are proposed.The following are the main contents of this paper.A graph based ranking algorithm is proposed.The bootstrapping relation extraction model can extract entity pairs for target relation.In order to improve the performance of the model,this paper proposes a graph based ranking algorithm to improve the result of the model,which takes into account the similarity between the result and the seed set.A model of seed set construction for target relation is proposed.The traditional method requires a lot of manual intervention,which would reduce the efficiency of experiment.This method constructs a Chinese semantic knowledge base based on Baidu encyclopedia,then classify the relations in knowledge base.This paper only consider three kinds of relationship,finally using knowledge base and search engine to construct seed set.The method for entity pairs similarity is improved.In graph based ranking algorithm,it is important to compute the similarity between pairs of entities.This paper use the potential relation analysis(LRA)to calculate the similarity,which can solve the problem of dimension and noise.The method for content patterns similarity is improved.In the graph based ranking algorithm,it is necessary to construct a content pattern diagram,and the similarity computation between content patterns is very important.In this paper,content pattern is represented by the Path-enclosed Tree.This paper uses the convolution tree kernel function method to calculate the similarity between content patterns.Finally,this paper constructs a visual character relationship map.The result of experiment proves that the feasibility and applicability of the relation extraction model,which can be used to extract any type of relationship.