Statistical Analysis
The pH and TTA (response variables) obtained during fermentation were subjected to regression analysis and analysis of variance (ANOVA) to determining regression coefficients and statistical significances of model terms and to fit the mathematical models to the experimental data, aiming at an overall optimal region for the response variable, multiple regression coefficients were determined by employing the least-squares technique to predict linear and quadratic polynomial model for the response variable studied. The behavior of the response surface was investigated for the response function (Y, the predicted response) using the regression polynomial equations. The generalize polynomial model proposed for the predicting the response variable is given as
Y=βo1x12x2+ β3x3 + β12x1 x2+ β13x1 x3+ β23x2 x3+ β11x21 + β22x22 + β33x23
Where βo is the intercept, β12 and β3 as coefficient. The significance of the equation parameters for each response variable was also assessed by F ratio at a probability (P) of 0.05. The adequacy of the models was determined using model analysis, lack of fit test and coefficient of determination (R2) analysis as describe by [8, 9], for a good fit of a model R2 should be at least 0.80 [10, 11]. The experimental design matrix, data and analysis, and optimization procedure were performed using the Design-Expert version 7 (state-Ease, Inc, Minneapolis, MN, USA).
3. RESULTS AND DISCUSSION