The Research Of The Data Mining Technique In The Audit Of The Basic Medical Security System For Employees

The basic medical security system for employees is an important part of our social security system,which is related to the stability and development of society.With the continuous improvement of our health care system and the incremental expansion of medical insurance,multi-level and wide medical security system is providing better medical resources and services for citizens.At the same time,the globalization of information has generated massive medical data,which includes not only the most critical financial information but also unutilized patient information,medical institution information and prescription information.On the basis of the introduction of data mining,data mining audit and relevant research status at home and abroad,this thesis clarifies the purpose and significance of the research on data mining audit of the basic employees' medical insurance data.The process of data mining audit is also exhibited in this dissertation by using SQL server 2008 R2 and SPSS Modeler software,through the example of the basic employees' medical insurance data audit in C city.In the meantime,based on the basic analysis of the data,the rules contained in basic medical insurance data are obtained and the characteristics of insurance frauds are summarized with CHAID decision tree and K-Means clustering model.After using data mining in the audit,audit risk is greatly reduced,while audit quality and efficiency are highly improved.According to the research results,this paper also puts forward some suggestions on improving medical data,strengthening the management of medical institutions and improving the fairness and safety of medical services.Furthermore,the data flow and data model which can be reused to audit the basic medical insurance for employees are acquired,and the advantages of data mining audit over the traditional audit mode are summarized.The main points of this thesis are as following:1.Introducing medical insurance data mining and the research status of data mining audit at home and abroad,and then pointing out the necessity of the research.2.Listing internal control audit methods and data audit methods,which are adopted in this dissertation.3.Investigating the existing internal control system of basic medical insurance for employees and drawing the relevant flowcharts.4.Collecting data from relevant units and importing it into the SQL server 2008 R2 software to complete the cleaning and sorting of data.Then the audit intermediate table is generated.5.Using the SPSS Modeler software for the data audit,drawing and running the relevant data flow according to the audit objectives,and completing the data analysis with the help of the CHAID decision tree and the K-Means clustering model.Finally,the audit suspects are determined and the audit conclusions are drawn with the model analysis results.6.Comprehensively evaluating and analyzing the audit results.