摘要
CUSUM控制图可用来监测均值的偏移。在单病种成本值发生重大差异前,短期内会发生一定失控趋势,CUSUM控制图却不能及时发出失控信号。针对成本差异分析中存在的问题,提出了基于支持向量机(SVM)的智能化单病种成本差异分析模型来替代CUSUM控制图。首先利用支持向量机(SVM)自身良好的泛化能力经过训练后获得复杂成本值之间的内在规律,然后对测试样本进行有效的预测分类,及时给出失控的样本点。将模型应用到单纯性阑尾炎病人样本中,与CUSUM控制图方法作比较,结果显示了该模型比CUSUM控制图更加及时发现有失控趋势或失控的样本点。
CUSUM control chart can be used to detect the shifts in the mean. Before the significant cost variance of single disease,the cost will have certain out-control trend in the short term,but the CUSUM control chart couldn＇t timely forecast. In view of the problems existing in the cost difference analysis, this paper put forward an intelligent method based on support vector machine（SVM） of single disease cost variance analysis model as an alternative approach to CUSUM control chart. Support vector machine（SVM） used its good generalization ability after training to obtain the internal rules from complex cost data, then the test sample points for effective prediction classification and timely forecast points of out-control. The model was applied to the simple appendicitis surgery, by comparing the method with the CUSUM control chart model,the results suggest this model is more sensitive than CUSUM control chart
出处
《工业工程与管理》
CSSCI
北大核心
2012年第5期11-15,共5页
Industrial Engineering and Management
基金
国家自然科学基金资助项目(70871086)
