肝癌术后无瘤生存期的人工神经网络预测

The prediction of the survival time of hepatectomy patients with artificial neural network method

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 。

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.