Page 180 - Contributed Paper Session (CPS) - Volume 4

CPS2166 Divo Dharma Silalahi et al.
4. Conclusion
The study has shown the promising of wavelength selection using input
scaling method particularly with application on a high dimension dataset such
NIRS spectral data. The proposed method was also robust since it applied a
robust measure of central tendency and robust measure of scale in the cut-off
threshold calculation. The proposed mod-VIP-MCUVE method has confirmed
the superiority to the other reference method such the conventional PLS with
no wavelength selection and input scaling applied, the VIP method, and the
MCUVE. In the selection of relevant wavelengths, using the modified cut-off
threshold the proposed method succeed to remove only the most irrelevant
wavelengths in the model, hence it also can still maintained the use of less
number of latent variables in the PLS model. Moreover, the proposed method
has confirmed the importance of wavelength selection method to reduce the
data dimension and to improve the model interpretability particularly to
investigate the fundamental attribute of diffuse selected reflectance of NIRS
spectral absorption and understanding of the system studied.

References
1. Centner, V., Massart, D. L., de Noord, O. E., de Jong, S., Vandeginste, B.
M., & Sterna, C. (1996). Elimination of uninformative variables for
multivariate calibration. Analytical chemistry, 68(21), 3851-3858.
2. Galindo‐Prieto, B., Eriksson, L., & Trygg, J. (2014). Variable influence on
projection (VIP) for orthogonal projections to latent structures (OPLS).
Journal of Chemometrics, 28(8), 623-632.
3. Kim, J., Kiss, B., & Lee, D. (2016). An adaptive unscented Kalman filtering
approach using selective scaling. In 2016 IEEE International Conference
on Systems, Man, and Cybernetics (SMC) (pp. 000784-000789). IEEE.
4. Kim, S., Kano, M., Nakagawa, H., & Hasebe, S. (2015). Input variable
scaling for statistical modeling. Computers & Chemical Engineering, 74,
59-65.
5. Natrella, M. G. (1963). Experimental Statistics Handbook 91. US
Government Printing Office.
6. Oussama, A., Elabadi, F., Platikanov, S., Kzaiber, F., & Tauler, R. (2012).
Detection of olive oil adulteration using FT‐IR spectroscopy and PLS with
variable importance of projection (VIP) scores. Journal of the American
Oil Chemists' Society, 89(10), 1807-1812.
7. Stuart, B. (2004). Infrared spectroscopy : fundamentals and applications.
Wiley: Canada, 78.
8. Trautmann, H. (2004). The desirability index as an instrument for
multivariate process control (No. 2004, 43). Technical Report/Universität
Dortmund, SFB 475 Komplexitätsreduktion in Multivariaten
Datenstrukturen.

169 | I S I W S C 2 0 1 9