Classifier for chinese traditional medicine with high-dimensional and small sample-size data

The identification of Chinese traditional medicine is a difficult subject in pharmacology. The development of chemical measurement and pattern recognition make chemical pattern recognition possible. In the paper a new chemical pattern recognition method is proposed, in which a simple method called corresponding-peak distance calculation is used to compute the distance between samples for a nearest neighbor (NN) classifier, and a genetic algorithm is used to optimize the parameters of the NN classifier. With the proposed method, experiments are carried out on chromatogram data of Panax. The results indicate that the method can identify the medicine material of different harvest time or habitats, furthermore, this method which combines pattern matching, genetic algorithm and NN classifier is robust, accurate and easy to implement.