Регрессионная зависимость числа падения аглютеновой мучной смеси от состава компонентов

Abstract

Modern technologies for the production of gluten-free flour products founded on flour mixtures are based on the processing of raw materials of plant origin, mainly flour with a high content of starch polysaccharides. A wide range of gluten-free bakery mixes has been developed, formulations for semi-finished and finished products have been designed by optimizing the chemical composition and nutritional value. The solution of the technological problem provides for multidimensional modeling of the relationship of a certain tightness between the studied quantitative characteristics by means of regression analysis. The regression dependence was studied using computer statistical analysis in the STATICTICA program by obtaining a Piece linear regression model with user-defined break-point. The break point (250 s) corresponded to the lower limit of the falling number for baking flour with normal autolytic activity. Regression analysis of the falling number of a gluten-free flour mixture with a break point made it possible to obtain two analytical linear dependencies for predicting the parameter values in separate areas up to the break point of the function. The constructed regression model is distinguished by high prediction accuracy (the multiple correlation coefficient R = 0,99981) and adequacy when the model residuals correspond to a normal distribution. The optimal amount of components in a gluten-free flour mixture with nor-mal autolytic activity (at 330 c in the experiment) was 20 % CMGP and 80 % ACM.