Discrete Wavelet Transform-Support Vector Regression Model and Its Application in NIR Analysis of Corn

摘要

Support vector machine (SVM) has become more and more popular as method for learning from examples. The basic theory is well understood. SVM is based on the principle of structural risk minimization, which makes it's generalization ability better than other traditional learning machine methods. Support vector regression (SVR) is based on SVM. In this paper, Discrete wavelet transform (DWT) combined with SVR was used in Near-infrared Spectroscopy analysis of corn. The purpose of this paper is to investigate the feasibility of DWT-SVM method in NIR analysis. Results suggest that the DWT-SVR model has better accuracy in forecast and higher computing speed than traditional methods.