Spectroscopy and Spectral Analysis, Volume. 32, Issue 1, 175(2012)
Construction and Simplification of the Calibration Model for Spectral Analysis of Fuel Oil Properties Based on Mutual Information Method
Near infrared diffuse reflectance (NIRS) and ultraviolet (UV) spectral analysis were adopted for quantitative determination of octane number and monoaromatics in fuel oil. Partial least squares regression (PLSR) was used for construction of vibrational spectral calibration models. Variables selection strategy based on mutual information (MI) theory was introduced to optimize the models for improving the precision and reducing the complexity. The results indicate that MI-PLSR method can effectively improve the predictive ability of the models and simplify them. For octane number models, the root mean square error of prediction (RMSEP) and the number of calibration variables were reduced from 0.288 and 401 to 0.111 and 112, respectively, and correlation coefficient (R) was improved from 0.985 to 0.998. For monoaromatics models, RMSEP and the number of calibration variables were reduced from 0.753 and 572 to 0.478 and 37, respectively, and R was improved from 0.996 to 0.998. Vibrational spectral analysis combined with MI-PLSR method can be used for quantitative analysis of fuel oil properties, and improve the cost-effectiveness.
Get Citation
Copy Citation Text
HAO Yong, SUN Xu-dong, CAI Li-jun, LIU Yan-de. Construction and Simplification of the Calibration Model for Spectral Analysis of Fuel Oil Properties Based on Mutual Information Method[J]. Spectroscopy and Spectral Analysis, 2012, 32(1): 175
Received: May. 11, 2011
Accepted: --
Published Online: Feb. 20, 2012
The Author Email: Yong HAO (haonm@163.com)