首页 | 本学科首页   官方微博 | 高级检索  
     检索      

肝癌术后无瘤生存期的人工神经网络预测
引用本文:贺佳,张智坚,贺宪民.肝癌术后无瘤生存期的人工神经网络预测[J].数理统计与管理,2002,21(4):14-16,27.
作者姓名:贺佳  张智坚  贺宪民
作者单位:1. 上海第二军医大学卫生统计学教研室,上海200433
2. 上海东方肝胆外科医院,上海,200438
基金项目:国家自然科学基金资助项目 (39770 835 )
摘    要:本文采用BP人工神经网络方法 ,构建预测肝癌患者生存期的网络模型 ,预测肝癌患者术后无瘤生存时间。该网络的回代贡献率为 83.94 % ,预测值与实际值的相关系数为 0 .916 2 ,误差均方为 0 .1347;网络对于检验集的贡献率为 71.11% ,预测值与实际值的相关系数为 0 .84 33,误差均方为 0 .16 0 8,生存期的预测值和实际值间比较 ,经配对t检验 ,t为 0 .3977,p为 0 .6 916。故认为BP神经网络预测肝癌患者的生存期效果较好 ,可进一步推广应用

关 键 词:BP神经网络  预测  肝癌  生存期
文章编号:1002-1566(2002)04-0014-04

The prediction of the survival time of hepatectomy patients with artificial neural network method
HE Jia ,HE Xian min ,ZHANG Zhi jian.The prediction of the survival time of hepatectomy patients with artificial neural network method[J].Application of Statistics and Management,2002,21(4):14-16,27.
Authors:HE Jia  HE Xian min  ZHANG Zhi jian
Institution:HE Jia 1,HE Xian min 2,ZHANG Zhi jian 1
Abstract:In this paper, the Bp artificial neural network method. was used to predict the survival time about hepatectomy patients. First, the network was set up and then to predict the survival time. For the training data, the contributing proportion of the predicting of the network is 83.94%, correlation coefficient between actual and predicting value is 0.9162, and the mean squared error for survival time is 0.1347, and for testing data, those are 71.11%, 0.8433 and 0.1608 respectively. The paired-t-test of actual and predicting survival time showed t=0.3977 and p=0.6916. So we can say, the BP neural network can be used to predict the survival time of hepatectomy patients and the resuld is better. It should be used further.
Keywords:BP neural network  predicting  liver cancer  survival time
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号