Using ANOVA-PCA for discriminant analysis: application to the study of mid-infrared spectra of carrageenan gels as a function of concentration and temperature.

In this work the ANOVA-PCA method is applied to a MIR spectroscopy dataset of carrageenan in order to evaluate which of the factors within its fixed effects experimental design are significant in relation to the residual error. The factors defined in the experimental design are concentration (1% and 2%), temperature (30, 40, 45, 50, and 60 degrees C), day (1 and 2) and sample (20 samples, 3 repetitions). The two factors, concentration and temperature, were considered as significant and the main features related with its physico-chemical properties were identified. It is also of interest to acquire a better understanding of the interaction between concentration and temperature and its effect on the adhesion of gels onto the surface of contact. In fact, no significant interaction was found between the two factors, but it was shown that the factor temperature behaves in a non-linear way. As classification using the ANOVA-PCA procedure has not been developed until now, a new method is proposed for the classification of new samples in respect to the levels of each significant factor.

[1]  Christelle Schaffer-Lequart,et al.  Characterization of commercial carrageenans by Fourier transform infrared spectroscopy using single-reflection attenuated total reflection. , 2004, Journal of agricultural and food chemistry.

[2]  António S. Barros,et al.  Influence of hydration of food additive polysaccharides on FT-IR spectra distinction , 2006 .

[3]  P. Mahalanobis On the generalized distance in statistics , 1936 .

[4]  Age K. Smilde,et al.  Statistical validation of megavariate effects in ASCA , 2007, BMC Bioinformatics.

[5]  Age K. Smilde,et al.  ANOVA-simultaneous component analysis (ASCA): a new tool for analyzing designed metabolomics data , 2005, Bioinform..

[6]  Peter de B. Harrington,et al.  Analysis of variance–principal component analysis: A soft tool for proteomic discovery , 2005 .

[7]  P. Pudney,et al.  Factors Associated with Dough Stickiness as Sensed by Attenuated Total Reflectance Infrared Spectroscopy , 2003 .

[8]  R. Barnes,et al.  Standard Normal Variate Transformation and De-Trending of Near-Infrared Diffuse Reflectance Spectra , 1989 .

[9]  P. Harrington,et al.  Proteomic analysis of amniotic fluids using analysis of variance-principal component analysis and fuzzy rule-building expert systems applied to matrix-assisted laser desorption/ionization mass spectrometry , 2006 .

[10]  J. J. Jansen,et al.  ASCA: analysis of multivariate data obtained from an experimental design , 2005 .

[11]  H. Hotelling A Generalized T Test and Measure of Multivariate Dispersion , 1951 .

[12]  J. Neyman Second Berkeley Symposium on Mathematical Statistics and Probability , 1951 .

[13]  D. Rutledge,et al.  Application of the ANOVA-PCA method to stability studies of reference materials. , 2007, Analytica chimica acta.

[14]  D. Massart,et al.  The Mahalanobis distance , 2000 .