摘要
目的:探讨使用BP神经网络进行肿瘤诊断的应用价值。通过收集的60例良性肿瘤特征向量和60例恶性肿瘤特征向量作为建立人工神经网络所需要的BP网络训练样本,然后使用这一网络对某种肿瘤的各项指标数据进行实验研究。结果:30例待测样本纳入本研究,29例样本得到了较理想的输出数据,从而获得了较高的诊断率。结论:多层前向神经网络在判断肿瘤性质方面有较高的应用价值。
Objective：To evaluate the value of BP neural network for tumor diagnosis. Methods：The eigenvectors of 60 group benign and 60 group malignant tumors were taken as the BP network training samples for the artificial neural network, then the artificial neural network was used to study the data of a tumor. Results：This study included 30 groups samples, 29 ones obtained ideal output data and get a higher rate of diagnosis. Conduslon：The multi-layer neural network has better applied value for tumor diagnosis.
出处
《医学影像学杂志》
2009年第8期978-980,共3页
Journal of Medical Imaging
关键词
BP神经网络
肿瘤诊断
多层前向网络
计算机应用
BP neural network
Tumor diagnosis
Network of multi-layers
Application of computer
