Canonical correlation: a multivariate technique to determine the contribution of various dependent and independent variables

Abstract

Composite flour chapaties were prepared by blending wheat, guar gum, lentil and chickpea flours in different proportions to evaluate their hypoglycaemic and hypocholesterolaemic worth because of their fibre content.Stepwise regression analysis revealed that dietary fibre and intestine length contributed negatively, while peroxide value (POV) contributed positively to the serum cholesterol and glucose. Likewise, for total chapati scores, dough stability and phytic acid showed inverse contribution, while folding ability, dough development time, dietary fibre, POV, crude fibre and texture performed positively. The canonical correlation was applied to develop a model between two groups of variables, where the first group comprised three dependent variables, i.e. total chapati scores, cholesterol and glucose and the second set comprised thirteen independent variables, i.e. crude fibre, POV, phytic acid, dietary fibre, dough development time, dough stability, feed intake, intestine length, low-density lipoprotein, triglycerides, albumin, texture and folding ability. The selection of independent variables was based on the significant chemical, rheological, sensory and efficacy characteristics in a multiple stepwise regression. The value of coefficient of determination for dietary fibre of the flour samples was observed to be 0.99749 indicating that 99.74% of the variation in dietary fibre can be explained by all other variables in the second set. However, 17.35% and 4.83% of variation in total chapati scores and cholesterol levels, respectively, can be explained by the entire variables in the first group, while 17.20% of variation in glucose can be explained by all other variables of the first group. The squared multiple regression of each variable in the first set with all the variables in the second set indicated that a 99.97% variation in total chapati scores, cholesterol and glucose is explained by the variables in the second set. The F-values were significant for total chapati scores, cholesterol and glucose. The results substantiated that all the variables in the second set are important to account for the significant variation for the variables tested in the first set. Canonical analysis revealed the existence of a high correlation between dependent variables and other independent variables.

Canonical correlations among chemical, physical and morphological characteristics of Xaraés palisadegrass under rotational grazing

“…For example, it can be valuable in plant breeding and screening, where canonical correlations can direct the material selection for variables that are too difficult or too expensive to measure in high quantities, by finding other variables that can be easily measured or that require less resources to quantify. Butt et al (2007), found effectiveness of canonical analysis as a multivariate technique to determine biochemical and quality parameters of dietary fiber. The analysis revealed high correlation between the quality characteristics and the other components.…”