A Novel Approach for Classification of Chinese Herbal Medicines Using Diffusion Maps

作者：Luo, Dehan^*; Chen, Huiqin; Yu, Hao; Sun, Yunlong

来源：International Journal of Pattern Recognition and Artificial Intelligence, 2015, 29(1): 1550003.

DOI：10.1142/S0218001415500032

• Chinese herbal medicines
• electronic nose (E-nose)
• diffusion maps
• linear discriminant analysis
• classification and identification

The quality of Chinese herbal medicines (CHMs) and the authenticity of drugs are directly related to the effect of treatment. This paper presents a method based on diffusion maps for the identification of pungent CHMs. An electronic nose (E-nose) was employed to collect the smell print of different groups of pungent CHMs with different kinds and different harvesting time. First, the collected odor data was mapped to a high-dimensional space, then using the diffusion maps algorithm, the low-dimensional manifold features were extended to the high-dimensional odor data. Finally, the linear discriminant analysis (LDA) algorithm was employed to implement a linear classifier. The results show that, the combination of Diffusion maps and LDA algorithm can well distinguish four different kinds of pungent CHMs and three different harvesting time of the same kind, provides a new way for the classification and identification of CHMs.

• 出版日期2015-2
• 单位广东工业大学