Statistical Analysis Of Medical Survival With Censored Data And Rough Set Decision Analysis

With the rapid development of computer technology, data mining methods continue to grow. Rough set theory and method are an effective analysis and processing inconsistent, inaccurate, incomplete information and other data analysis tools. Because rough set decision analysis method does not depend on expert knowledge and experience, is only dependent on the objectivity of the data itself, has been generally recognized by the statistics and other disciplines. In this paper, rough set data mining method conducted in-depth research, focusing on analysis based on rough set attribute reduction algorithm in data mining application of the rules extraction stage. It includes upper (lower) approximate relationship, knowledge reduction, core, indistinguishable relationship, rules extraction. Survival analysis is the study of the phenomenon of survival and response time data and their statistical regularities of a discipline. The disciplines in biology, medicine, insurance, science, reliability engineering, demography, sociology, economics and other aspects have important applications.In this paper, an actual medical data are diabetic Cox regression survival analysis and rough set of medical data aided rule extraction. The main features of survival analysis are they can handle censored data. Censored data are the exact time which are not observed. The objects involved in the paper are149people with diabetes’data, which are analyezed with SPSS statistical software and modules in the rough set of MATLAB respectively. The results from different aspects are derived. This paper considers a medical data table, which involves two methods of survival analysis and rough set through counting the data of survival indicators. Through the establishment of semi-parametric model in survival analysis methods, using SPSS software Cox regression function modules, based on survival time function, the interaction between hazard function and coefficient is modified so that the effectiveness of hazard function affected by coefficient can be checked. This method can evaluate factors effect without making assumptions about the specific distribution of survival time, greatly simplifying the process of solving survival analysis.By discretization and rough set method survival data can be divided into an indistinguishable relationship rough calculated. Knowledge reduction in rough set and the application of decision tables in knowledge representation is intruduced. At the same time, through studying the composed of rough set data analysis system and basic algorithm, exemplified system implementation procedures, including attribute reduction, core, and so on. Finally, we use an example to verify the correctness of the program, implement solution of Indistinguishable relationship, reduction, core directly by using the MATLAB program in rough set. By using rough set rules extraction, People with diabetes has been determined reduction of adjuvant therapy results. The actual results of the diagnosis and hospital programs fit very well. Results obtained from two broad approaches have overlapping and difference. Actual results verify the correctness proposed in the paper.