摘要
目前临床病例异常检测的研究主要采用病症关联、费用控制和临床序列模式挖掘等方法,对无症状信息、无完整临床行为时间等临床数据仍具有一定的局限性。根据这一类临床数据特点,提出了基于模式识别的CC_FR模型,该模型采用频繁模式挖掘的方法确定单病种隶属函数,通过隶属函数中的频繁模式与待检测临床病例相匹配得到检测结果。实验结果表明,该模型可以有效的检测临床病例异常性,在临床医疗中起到监督和警示的作用。
Existing anomaly detection of clinical case usually founded on association with symptom and disease,expenses control and pattern mining on clinical sequence.Some clinical data that no information of sumptom and time fall across some problem.Based on the characteristic of these data,CC_FR model is advanced,this model used frequent patterns mining to confirm subordinate function and matched frequent patterns of subordinate function with clinical data to gain the detection results.The experimental results show that the model can get better general performance,it can play the supervising and caution function in clinical medical treatment.
出处
《计算机工程与设计》
CSCD
北大核心
2009年第24期5705-5707,5711,共4页
Computer Engineering and Design
基金
江苏省高技术研究基金项目(BG2007028)
关键词
异常检测
模式识别
频繁模式挖掘
FP增长
隶属函数
anomaly detection
pattern recognition
frequent patterns mining
FP_Growth
subordinate function
