人工神经网络在中医证候学中的应用初探

Application of artifi cial neural network in traditional Chinese medicine syndromes

文章总结了人工神经网络在证候诊断中的优势、应用现状、存在的问题。运用其非线性映射特性,辨识中医四诊信息,利用自身学习能力,获取证候规律,建立非线性映射关系,建立证候诊断模型,为准确诊断新病例的证型提供技术支持。在训练过程中可能会出现学习速度慢、信息不全面和准确率低的情况,需要采用改进学习算法和增加样本含量等方法来解决。

The superiority, application status and problems of artificial neural networks were summarized in the study of TCM syndromes. With a strong nonlinear mapping, the recognition of four diagnostic information about TCM, to obtain the rule of TCM syndromes by its own learning ability, and to make a creation of the nonlinear relationship, establishment of syndrome diagnosis model in diagnosing new cases accurately which can provide a technical support. Artificial neural network might have some problems such as learning with slow rapid, incomprehensive information and low accuracy in the training process. The problem need to be solved by improving learning algorithm and increasing sample size.