A hybrid feature selection combining wavelet transform for quantitative analysis of heat value of coal using laser-induced breakdown spectroscopy

A hybrid feature selection method combining with wavelet transform (WT) was proposed to analyze the heat value of coal using laser-induced breakdown spectroscopy (LIBS). The hybrid feature selection method consisted of distance correlation (DC) method and recursive feature elimination with cross-validation (RFECV) method, which combined the advantages of DC-based filter method and RFECV-based wrapper method. First, WT method was used to filter noise signal from LIBS spectra of coal samples, and the de-noised wavelet coefficients were obtained. Second, the de-noised wavelet coefficients were further eliminated by the hybrid feature selection method. Finally, the retained wavelet coefficients were used directly as input variables to establish a prediction model for heat value determination of coal. 28 powdery coal samples were used in this experiment, of which 21 were calibration set and 7 were validation set. The effectiveness of the hybrid model was studied. Compared with several other models, the proposed hybrid model showed the greatest improvement in predictive accuracy and precision, and the computing time has been greatly reduced. The experimental results demonstrated that the hybrid model can effectively reduce the calculation time and improve the performance of the model.