摘要
在医保基金管理中,第三方付费机制和信息不对称等问题造成了基金运作面临严重"道德风险"困境,医药机构和参保人可能存在过度使用医保基金的倾向。通过对参保人就医行为序列的分析挖掘其就医行为模式,对于发现疾病发病规律、参保人健康状况以及是否存在违规欺诈行为,从而有效防范基金风险具有非常重要的作用。由于就医行为模式的特殊性,传统的序列模式挖掘算法在结果可用性和效率上存在问题,如挖掘结果丢失时间间隔较长的模式,挖掘过程需多次构造投影数据库等,因此难以直接应用。针对就医行为模式特点,提出了基于二叉树增长策略的向量模式挖掘算法VPM。实验表明,VPM算法在解决就医行为模式挖掘问题上具有良好的性能。
With regard to medical insurance fund management,problems deriving from third-party payment mechanism and imbalance of information acknowledgement are causing the plight of ＂immorality risk＂ along with the operation of the fund.Both medical agencies and patients tend to excessively exploit the medical insurance fund.Through analysis of medical treatment behaviour sequence of the insured people to mine their medical treatment behaviour pattern,it is easier to discover the regularities of diseases,the health status of the insured people,and whether there is any fraud behaviour,therefore it plays an important role in controlling that fund risk.Due to the special nature of the medical treatment behaviour pattern,there is deficiency on its result feasibility and efficiency,such as the missing of longer interval patterns in mining results and the requirement for repeatedly building shadow casting databases,etc.for traditional sequential pattern mining algorithms.Therefore their immediate application encounters barriers Aiming at the specialties of medical treatment behaviour pattern,the authors propose a novel algorithm based on binary-tree growth strategy,called Vector Pattern Mining algorithm（VPM）.Theoretical analysis and experimental results show that VPM algorithm perform well in solving the problem of medical treatment behaviour pattern mining.
出处
《计算机应用与软件》
CSCD
2011年第8期123-125,180,共4页
Computer Applications and Software
基金
上海市科委科研计划基金项目(08511500203)
上海市重点学科建设项目(B114)
关键词
医保基金
风险防控
就医行为模式
序列模式
数据挖掘
Medical insurance fund Risk prevention and control Medical treatment behaviour pattern Sequential pattern Data mining
