| Cox模型与BP神经网络在处理非线性数据时的性能比较 |
| |
| 引用本文： | 贺宪民,贺佳,范思昌.Cox模型与BP神经网络在处理非线性数据时的性能比较[J].数理统计与管理,2004,23(2):69-72. |
| |
| 作者姓名： | 贺宪民 贺佳 范思昌 |
| |
| 作者单位： | 第二军医大学卫生统计学教研室,上海,200433 |
| |
| 基金项目： | 国家自然科学基金资助项目(39770835) |
| |
| 摘 要： | 本文采用BP神经网络、Cox模型和bootstrap方法,比较BP神经网络与Cox模型在处理非线性资料时的性能。两种方法的预测一致性的均数分别为0.7525和0.7706。对于非线性资料,BP神经网络的预测效果优于Cox模型。
|
| 关 键 词： | BP神经网络 Cox比例风险模型 预测一致性 bootstrap |
| 文章编号： | 1002-1566(2004)02-0069-04 |
| 修稿时间： | 2002年6月21日 |
Comparison of nonlinearity disposing ability between Cox proportional hazards regression model and BP Neural Network |
| |
| HE Xian-min,HE Jia,FAN si-chang.Comparison of nonlinearity disposing ability between Cox proportional hazards regression model and BP Neural Network[J].Application of Statistics and Management,2004,23(2):69-72. |
| |
| Authors: | HE Xian-min HE Jia FAN si-chang |
| |
| Abstract: | To compare the nonlinearity disposing ability between BP neural network and Cox proportional hazards regression model, we studied BP neural network, Cox proportional hazards regression model and leave-one-out method. Means of C values of two methods are 0.75245 and 0.77055 respectively. The predictive concordance of BP neural net work is higher than Cox proportional hazards regression model. |
| |
| Keywords: | BP neural network Cox proportional hazards regression model Predictive concordance bootstrap method |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
