VPM:A MEDICAL TREATMENT BEHAVIOUR PATTERN MINING ALGORITHM

Zhou Kun,Wang Airong,Zhang Jingyi,Xiong Yun,Zhu Yangyong

DOI: https://doi.org/10.3969/j.issn.1000-386X.2011.08.036

2011-01-01

Abstract: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.