RBF网络在大肠癌患者术后存活期辅助预测中的应用

摘要

目的 探索径向基函数(RBF)网络在大肠癌患者术后存活期辅助预测中的应用.方法 利用RBF网络综合研究大肠癌患者的各项指标与术后存活期的关系.将随访收集的数据按照5.5∶1的比例随机分为训练集和测试集,建立用于大肠癌患者术后存活期辅助预测的RBF网络.结果 所建立的RBF网络对训练集数据的预测准确率为100%,对测试集数据也具有较高的预测准确率.结论 RBF网络可以仿真大肠癌患者输入指标与术后存活期之间的关系,并能克服逆传播(BP)网络的缺陷.%Objective To investigate the application of radial basis function (RBF) network in assisting the forecast of postoperative survival time in patients with colorectal cancer.Methods Using RBF network, the relationship between the indexes and postoperative survival time of colorectal cancer patients was studied.Data were collected from patients during follow-up visits, and randomly divided into a training set and a testing set at a ratio of 5.5∶1, in order to establish an RBF network forecasting postoperative survival time in patients with colorectal cancer.Results The established RBF network was applied for forecasting the data in the training set, with a precision of 100%.It could also be adopted to forecast the data in the testing set, with high precision.Conclusions RBF network can simulate the relationship between the input indexes and postoperative survival time of colorectal cancer patients, and overcome the disadvantages of BP network.