The Establishment Of The Chinese Herb And The Research In The Properties Of Anns

In recent years, Artificial Neural Network is used fairly widely in fields of pattern recognition and performance prediction of biological medicine, such as clinical prediction, computer assisted diagnosis, the identification and examination of medical signal(Neural Network is applied in the image recognition of medical examination), the classification of disease and modeling prediction. And Artificial Neural Network is comprehensive applied in drug analysis, Chinese herbal medicine identification, technical control of medicine production and so on. Medicine research is too complicated because of many phenomena in medicine filed possessing characteristic of nonlinear, dynamic, random and uncertain. Artificial Neural Network is applied in performance examination and prediction of cold Chinese traditional medicine in this paper. The study is combined with Neural Network for specific applications on the base of the analysis of existing herbal medicines performance prediction. The mainly work and conclusions of the thesis are as follows.1.Introducing the related applications and research status in the field of medicine of Artificial Neural Network, emphatically describing the background of pharmic research. 2.Introducing BP Neural Network algorithm, analysing several important aspects of BP Neural Network model design and establishing basis for BP Neural Network model applied to traditional medicine research.3.Make sure 20 kinds of drug property including(Four Properties, Five Tastes, Meridian, etc)is the main influencing factors to establish and analyse BP Neural Network by collecting and analysing a large of cold Chinese traditional medicine data.4.Proposing the forecasting models based on the principle that BP Neural Network can be applied to predict. And discussing the key link of Neural Network prediction model such as drug data sample selection and pre-processing, input and output data selection, making sure the number of node in hidden layers, the selection of activation function and training algorithm, etc.5.According to the results of medicines performance prediction and test by BP Neural Network model based on 60 cold Chinese traditional medicine data, Proving BP Neural Network can reveal the complex relationship between basic attributes (Four Properties, Five Tastes, Meridian, etc)and property of drugs in smaller error range.