摘要
目的:优化中医证候诊断模型,为中医证候诊断标准的研究提供可行性方法。方法:提出用于中医证候诊断的径向基(RadialBasisFunction,RBF)神经网络,利用聚类分析确定RBF神经网络隐层的参数,运用最小二乘确定RBF神经网络输出层的参数。结果:通过模型检验,证候诊断模型判准率比BP网络模型判准率高;证候诊断模型训练速度比BP网络模型快。结论:基于聚类分析的RBF神经网络用于中医证候诊断的研究是可行的和有效的。
Objective:optimizing the model of TCM syndrome diagnosis, providing a feasible method in application of study of TCM syndrome diagnosis criterion. Methods:presenting RBF Neural Network based on clustering analysis in the application of TCM syndrome diagnosis, defineing the parameters of the hidden layer by means of clustering analysis, determining the arguments of the linear output layer by way of Least Squares Algorithm,establishing the model of RBF Neural Network based on clustering analysis in the application of TCM syndrome diagnosis, Result:simulation results suggest that TCM syndrome diagnosis model in this paper has higher modeling accuracy than the original BP network, and the learning speed of TCM syndrome diagnosis model in this paper converges much faster than that of the original BP network. Conclusion : It is practical and valid for T-S model in this paper to be applied to the study of TCM syndrome diagnosis.
出处
《中国中医基础医学杂志》
CAS
CSCD
北大核心
2005年第9期685-687,共3页
JOURNAL OF BASIC CHINESE MEDICINE
关键词
中医证候诊断
聚类分析
RBF神经网络
最小二乘法
TCM syndrome diagosis
clustering analysis
RBF neural network
Least Squares Algorithm