Discrimination of plant samples using near-infrared spectroscopy with a principal component accumulation method

Abstract

Discrimination is a universal problem in various fields. Near-infrared (NIR) spectroscopy combined with pattern recognition technique has shown great power and gained wide acceptance for analyzing complex samples. A principal component accumulation (PCAcc) method was proposed in our previous work for the two-class problem of cancer classification based on microarray gene expression data. In this work, PCAcc is applied to the multiclass problem of plant sample classification based on NIR spectroscopy. A hierarchical way is adopted as in the decision tree methods, each node dividing the samples into two classes. A parameter of resolution (R_s) is used to quantitatively measure the difference between the two classes. To validate the performance of the proposed method, it is applied to the NIR spectral datasets of different parts of tobacco leaves and different brands of cigarettes. The results show that the method can discriminate these samples with a very good performance in terms of classification accuracy.