The Study On Related Issues Of Research Fields Based On Text Mining

With the development of science and technology, the traditional natural science has been specialized; new cross-disciplines have come up; the trend of integration among research fields has gradually increased; and the development of research fields has become more comprehensive. As a result, it is critical for scientific researchers to explore the structure of research fields, to mine the research fronts, to analyze the intercross, the trends of integration and division of research fields, which will help them to know the status of disciplines, carry out in-depth researches and promote the development of science.Based on the comparison and analysis on different methods of research field analysis, text mining is chose to divide the research fields by mining the proposals of basic researches. The status, front focus, overlapping and evolution of research fields of management science and engineering has been analyzed from two aspects of subject codes and subject clusters as following:(1) Divide the research fields. The corpus has been processed using text classification and text clustering to determine the scope of research fields. Centroid classifier is applied and hierarchy-based clustering is also used.(2) Analyze the status of research fields and identify the front focus. Firstly the research methods of identifying the front focus are introduced. Then the status and front focus of subject codes and subject clusters are analyzed. Bibliometric method is used to identify the focus fields in management science and engineering, while keyword analysis, is applied to identify the focus topics of knowledge management.(3) Analyze the overlapping of subject codes. Three variables are defined as code clarity, code ambiguity and code overlapping to explore the intercross status of management science and engineering, especially the overlapping situation of management system engineering.(4) Analyze the evolution of research fields. The relationships of evolution and traceability are defined. According to the network of maximal similarity between adjacent subject clusters, the trends of evolution of subject clusters from 2000 to 2005 are analyzed. And using domain keywords, the evolution of detailed research topics of knowledge management is analyzed from two aspects of the subject codes and subject clusters.The results tell that the structure of management science and engineering is complicated and finance, knowledge management and etc. are front focus. With the aggravation of the intercross and ambiguity of subject codes year by year, the trend of evolution is clear. The results of this paper will help scientific researchers to know about the status and trend of disciplines.