Study on Compound Prescriptions Pattern Recognition Based on Fuzzy Neural Network

Abstract

CA new method of Chinese patent medicine pattern recognition based on Neural Network and prescriptions Expert System is presented. In this method, first, chromatogram experiments are done on a prescriptions preparation at uniform condition and then, peak separation of each group is separated and analyzed. Some possible traditional Chinese medicines are recognized over the use of fuzzy neural network. Prescriptions Expert System proposes the most possible medicinal herbs needed to be further recognized on the basis of the recognized medicinal herbs set and credibility of the possible medicinal herbs outputted by fuzzy neural network. Chromatogram experiments on the basis of specificity of are done on the most possible medicinal herbs in order to verify whether the possible medicinal herbs are the expected ones according to specificity index, the proportion of the medicinal herbs in this prescription can be calculated simultaneously. By circulating above-mentioned method, all the assembled medicinal herbs and their proportions in compounded medicinal herbs can be recognized and consequently prescriptions are recognized. The efficiency and accuracy of this method is verified by actual compound preparation testing. Furthermore, the method is shown to be able to provide a new approach to analysis of compound prescriptions.