Indian Journal of Science and Technology

Open Access Open Access  Restricted Access Subscription Access

A Comparative Study of Principal Component Regression and Partial least Squares Regression with Application to FTIR Diabetes Data


Affiliations
1 Department of Statistics, Tuberculosis Research Centre, ICMR, Chennai-600 031, India
2 P. G. Department of Mathematics, Pachaiyappa’s College, Chennai-600 030, India
3 P. G. Department of Physics, Pachaiyappa’s College, Chennai-600 030
 

In recent years, Fourier Transform Infrared (FT-IR) spectroscopy has had an increasingly important role in the field of pathology and diagnosis of disease states. The principal component regression (PCR) and the partial least squares regression (PLS) are the often proposed methods and widely used in FTIR data analysis, when the number of explanatory variable is relatively large in comparison to the samples as the least squares estimator may fail in such situations. They provide biased estimators with the relatively smaller variation than the variance of the least squares estimators. In this paper, a FTIR diabetes dataset is used in order to examine the performance of the two biased regression models on prediction. The conclusion is that for prediction PCR and PLS provides similar results which require substantial verification for any claims as to the superiority of any of the two biased regression methods.

Keywords

Fourier Transform Infrared, Principal Component Regression, Partial least Square, Diabetes Data
User

  • Arnold MA and Small GW (1990) Determination of physiological levels of glucose in an aqueous matrix with digitally filtered Fourier Transform Near-Infrared Spectra. Anal. Chem. 62, 1457-1464.

  • Beebe KR and Kowalski BR (1987) Comparison of multivariate calibration and analysis. Anal. Chem. 59. 1007A-1017A.

  • Bertacche V, Pini E, Stradi R and Stratta F (2006) Quantitative determination of amorphous cyclosporine in crystalline cyclosporine samples by fourier transform infrared spectroscopy. J. Pharma. Sci. 95, 158-166.

  • De Jong S (1993) SIMPLS An alternate approach to PLS regression. Chemometrics & Intelligent Lab. Systems. 18: 251-263.

  • De Jong S and Kiers H (1992) Principal convent regression chemometrics and intelligent laboratory systems. 14, 155-164.

  • Geladi P and Kowlakski BR (1986) Partial least squares methods regression: A Tutorial Anal. Chemica Acta. 185, 1-17.

  • Griffiths PR and de Haseth JA (1986) Fourier transform infrared spectrometry. Wiley, NY.

  • Haaland DM and Thomas EV (1988a) Partial least- squares methods for spectral analyses. 1. Relation to other quantitative calibration methods and the extraction of qualitative information. Anal. Chem. 60, 1193-1202.

  • Haaland DM and Thomas EV (1988b) Partial least- squares methods for spectral analyses. 2. Application to simulated and glass spectral data. Anal. Chem. 60, 1202-1208.

  • Kartheek M, Anton Smith A, Kottai Muthu A and Manavalan R (2011) Determination of adulterants in food: A review. J. Chem. & Pharma. Res. 3(2), 629- 636.

  • King H, Aubert RE and Herman WH (1998) Global burden of diabetes,1995-2025. Prevalence, numerical estimates and projections. Diabetes Care. 21, 1413- 1414.

  • Low-Ying S, Shaw R A, Leroux M and Mantsch HH (2002) Quantification of glucose and urea in the whole blood by mid-infrared spectroscope of dry films. Vibr. Spectr. 28,111-116.

  • Marquardt LA, Arnold MA and Small GW (1993) Near-infrared spectroscopic measurement of glucose in a P-protein matrix. Anal. Chem. 62, 3271-3278.

  • Nestor PG, O’Donnell BF, McCarley RW, Niznikiewiez M, Barnard J, Jen S Z and Bookstein FL (2002) A new statistical method for testing hypotheses of neuropsychological/ MRI relationships in schizophrenia; Partial least squares analysis. Schizophrenia Res. 53, 57-66.

  • Shaw RA, Kotowich S, Leroux M and Mantsch HH (1998) Multianalyte serum analysis using mid-infra spectroscopy. Ann. Clinical Biochem. 35, 624-632.

  • Trygg J (2004) Prediction and spectral profile estimation in multivariate calibration. J. Chemometrics. 18,166–172.

  • Wold S, Sjo¨stro¨m M and Eriksson L (2001) PLS- regression: A basic tool of chemometrics. Chemometrics & Intelligent lab. Sys. 58, 109-130.

  • Zhou X, Hines P and Borer (1998) Moisture determination in hygrosckoic drug substances by near infrared spectroscopy. J. Pharma & Biomed. Anal. 17(2), 219-225.


Abstract Views: 848

PDF Views: 165




  • A Comparative Study of Principal Component Regression and Partial least Squares Regression with Application to FTIR Diabetes Data

Abstract Views: 848  |  PDF Views: 165

Authors

P. Venkatesan
Department of Statistics, Tuberculosis Research Centre, ICMR, Chennai-600 031, India
C. Dharuman
P. G. Department of Mathematics, Pachaiyappa’s College, Chennai-600 030, India
S. Gunasekaran
P. G. Department of Physics, Pachaiyappa’s College, Chennai-600 030

Abstract


In recent years, Fourier Transform Infrared (FT-IR) spectroscopy has had an increasingly important role in the field of pathology and diagnosis of disease states. The principal component regression (PCR) and the partial least squares regression (PLS) are the often proposed methods and widely used in FTIR data analysis, when the number of explanatory variable is relatively large in comparison to the samples as the least squares estimator may fail in such situations. They provide biased estimators with the relatively smaller variation than the variance of the least squares estimators. In this paper, a FTIR diabetes dataset is used in order to examine the performance of the two biased regression models on prediction. The conclusion is that for prediction PCR and PLS provides similar results which require substantial verification for any claims as to the superiority of any of the two biased regression methods.

Keywords


Fourier Transform Infrared, Principal Component Regression, Partial least Square, Diabetes Data

References





DOI: https://doi.org/10.17485/ijst%2F2011%2Fv4i7%2F30103