Research On Named Entity Recognition And Visualization For Social Dynamic Information Analysis

Within the era of big data,computer and information processing technology provide efficient support on the improvement of military information systems.At present,it still stays on manual processing for the armed forces to deal with those social aspects of dynamic information analysis and processing,which leads to poor efficiency in decision making for commanders.As a result,the information extraction technology arises at the historic moment,and the named entity recognition method can solve this kind of problem.As a result,the information extraction technology came into being,named entity recognition as the basic research work of information extraction technology,has been widely concerned.It can be used to not only improve the work efficiency,but also optimize the statistical analysis.This paper mainly studies on the named entity recognition technology for the dynamic information of the society,which includes three parts:1.Evaluation and research on Chinese named entity recognition tool.Named entity recognition based on the current three main open source tools LTP,HanLP,BosonNLP are evaluated respectively based on 1998 ”People's Daily” corpus for the performance evaluation and comparison.The results show that BosonNLP has the best performance in the social dynamic information text,and the F factor is 90.27%.2.Named entity recognition method based on feature clustering and CRF.According to the three current popular open source named entity recognition tools,a new model based on feature clustering combined with CRF is put forward for named entity recognition.Compared with BosonNLP,the proposed method can achieve better performance,including 5.41% of precision,3.01% of recall and 4.21% of F factor for personal name recognition,and 3.79% of precision,2.87% of recall and 3.34% of F factor for place name recognition.3.Dynamic information visualization based on digital map API.According to the named entity recognition results,the social dynamic information can be displayed online through three aspects separately,which include time dimension,area dimension,and events dimension.