Three Modeling Methods Application in Near-Infrared Spectra Analysis for Determination of Volatile in Lignite Coal Samples

Abstract

We studied volatile determination in lignite coal samples using near-infrared (NIR) spectra. Firstly, spectra were pre-processed to eliminate useless information. Then, determination model was constructed by partial least squares regression. We used discrete wavelet transform to pre-processing. To study the influence of modeling on determination of volatile for NIR analysis of lignite coal samples, we applied three techniques to build determination model, including support vector regression, partial least square regression and radial basis function neural network. Comparison of the mean absolute percentage error (MAPE) and root mean square error of prediction (RMSEP) of the models show that the models constructed with radial basis function neural network gave the best results.