Classification of traditional Chinese medicine by nearest-neighbour classifier and genetic algorithm

Zhang Lixin1, Zhao Yannan1, Yang Zehong1, Wang Jiaxin1, Cai Shaoqing2, Liu Hongyu2
1The State Key Laboratory of Intelligent Technology and Systems Computer Science & Technology Department, Tsinghua University, Beijing, China
2Pharmacy Department Medicine College, Peking University, Beijing, China

Tóm tắt

The identification of traditional Chinese medicine is a difficult subject in pharmacology. The development of chemical measurement and pattern recognition make chemical pattern recognition possible. In this paper a new chemical pattern recognition method named NN2GA is proposed, in which a simple method named 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. A method named NN3GA, which is realized by adding a parameter to NN2GA, is proposed to improve the performance of the classifier. Experiments are carried out on chromatogram data of Panax, and comparisons are made between NNPR, NN2GA, and NN3GA classifiers. The results indicate that the method which combines NN with a genetic algorithm can identify medicine material having different harvest times or habitats. Furthermore, this method is robust, accurate and easy to implement.

Từ khóa

#Genetic algorithms #Neural networks #Nearest neighbor searches #Testing #Chemical technology #Optimization methods #Chemical analysis #Principal component analysis #Laboratories #Intelligent systems

Tài liệu tham khảo

yuzhu, 1991, Quality assessment of the chinese traditional medicine rhubarb by chemical pattern recognition, Acta Pharmaceutica Sinica, 26, 132 yanjiang, 1995, Application of artificial neural networks to the feature extraction in chemical pattern recognition of the traditional chinese medicine venenum bufonis, Acta Pharmaceutica Sinica, 30, 698 xu, 1993, Quality assessment of the traditional chinese medicine by chemical pattern recognition, Acta Pharmaceutica Sinica, 28, 936 chen, 1996, Genetic algorithm and its application, People's Posts & Telecommunications Publishing House goldberg, 1989, Genetiv Algorithm as Search Optimiation and Machine Learning weijia, 1996, Matching for unknown crude herb by gc-computer, Journal of China-Japan Friendship Hospital, 10 qianguan, 1999, Studies on the quality of american cinseng(panax quinque folius ) by chemical pattern recognition, Chinese Traditional and Herbal Drugs, 30 xiukun, 1996, Studies on radix sophorae flayescentis quality by chemical pattern recognition, China Journal of Chinese Materia Medica, 21 kohavi, 0, The Wrapper Approach in Feature Selection f�r Knowledge Discovery and Data Mining, 33 davis, 1991, Handbook of Genetic Algorithms keinosuke, 0, Introduction to statistical pattern recognition, Boston R Academic, 219 aha, 1991, Instance-Based Learning Algorithms, 6, 37