A method for identification of moxa grade by Fourier transform infrared spectroscopy fingerprint was proposed,and the optimal model for identifying moxa grade was obtained by comparison of combination of 8 spectral preprocessing methods[denoising,Gaussian filtering,multivariate scattering correction,standard normal transformation,first derivative+Savitzky-Golay(SG)smoothing,second derivative+SG smoothing,first derivative+Norris Gap,and second derivative+Norris Gap]and 5 pattern recognition methods[back propagation neural network(BP-NN)algorithm,genetic optimization support vector machine(SVM-ga),particle swarm optimization support vector machine(SVM-pso),random forest(RF)algorithm,and K-nearest neighbor(KNN)algorithm].As shown by the results,there were 11 common peaks in maxo fingerprint.Nine principal components were obtained by principal component analysis,and the cumulative variance contribution rate reached 99.67%.The combination of standard normal transformation and SVM-pso algorithm had the best discrimination effect,with the discrimination accuracy of 100%in the training set and 93.3%in the test set.
维普
万方数据