A Research Of Topic Evolution Detecting About Medical Field Based On Social Network Analysis

With the increase in the amounts and types of medical information resources and the interdisciplinarity of the related works,it has become increasingly challenging for researchers and information personnel to grasp the theme development.How to recognize the evolution of research topics in massive scientific research information using advanced computer technology has become a key problem in present day scientific research.Aiming at solving the problem,this paper attempts to discover an effective method to automatically identify the topic evolution in a specific subject area,so as to better facilitate researchers and information personnel in their work.In this paper,we first analyze related theories and technical approches of topic evolution identification at home and abroad,we emphatically analyze the characteristics of medical topics evolution and the evolution analysis method based on social network analysis,then we summarize its enlightenment for topic evolution judgment and the current problems existed in the practical application.Considering the prominent position of medical research among all the subject areas in scientific research,and in order to carry out a new topic evolution detection method,the author proposes a model based on social network analysis for judging topic evolution in medical researches and demonstrates its operating process.The main stages in the process include medical words extraction,topic area identification,topic association,the identification of topic inheritance,freshman and death events and the identification of topic spliting and merging events,the identification of topic spliting and merging events includes key topic identification,the identification of the main path of key topic and the spliting and merging events on the main path.On the medical words extraction,this paper maps the UMLS words using the text mapping tool MetaMap and designs the topic extraction scheme based on semantic type,we adopt the way of 'word-semantic type' to show the text content.We use LDA model to identify the topic which is showed with a group of closely related keywords set.On topic correlation,we use cosine similarity calculation method,and we identify key topic using point center degrees,we adopt the way of the sum of the text similarity computing key topic evolution main path,and with the help of the related algorithm to identify the events of spliting and merging.This paper continues to take the study of breast cancer disease treatment research as a field to test the new model for identifying topic evolution in medical research.The test results are highly concordant with authoritative literature reviews in the field and are further confirmed by interviews with the field's leading experts,thus verifying the reliability of the techniques and approaches proposed by the study.With the above efforts,this study designs and realizes the model for detecting topic evolution in medical research,in the process of which an effective method for indentifying topic evolution in the medical research field is discovered,and will have positive effect in facilitating the identification and measurement of topic evolution in information analysis.