数据挖掘的流程改进和模型应用

Process Optimization and Model Application of Data Mining

针对跨行业数据挖掘标准流程(CRISP-DM)的不足,提出在结果部署和数据理解之间增加连线,从而形成了数据挖掘过程的回环.研究了模型部署的方法,用数据库和批处理文件等技术解决了模型应用于信息系统中的平台限制.以陕西省高校财务风险预警为例,在Clementine上运用支持向量机作为挖掘算法,按照改进的流程进行了实证.改进的数据挖掘流程和实现方法对于数据挖掘模型的广泛使用具有推动作用.

This paper addresses the shortage of CRISP-DM（cross-industry standard process for data mining）,propose add connection between evaluation and deployment to form the circle of data mining process.Based on the research of development method,the technologies of database and batch file is used to overcome the platform limitation of model in information system.According to the optimized process of data mining,early warning of financial risks of universities in Shaanxi province is implemented with support vector machines in Clementine.Improved data mining process and realization of the data mining model are useful to widen the data mining model.