A = Mass of root; B = Fermentation time; C= Fermentation temperature
The experimental equation deduced for fermentation of cassava root for
‘pupuru ’ production is given by:
\begin{equation}
\text{pH}=4-0.020A-0.35B-0.04C-0.037\text{AB}+0.038\text{AC}-0.012\text{BC}-0.045A^{2}+0.2B^{2}+0.055C\nonumber \\
\end{equation}\begin{equation}
\text{TTA}=2.19+0.039A+0.064B+0.29C-0.11\text{AC}-0.15\text{BC}+0.22A^{2}+0.45B^{2}-0.051C\nonumber \\
\end{equation}The analysis of variance (ANOVA) of the regression parameters for the
response surface model is presented in Table 3. The ANOVA obtained
showed that the regression model obtained was significant (P <
0.05) for the response variable with R2 = 0.9674 and
insignificant lack of fit (P = 0.05). The goodness of fit of the
mathematic model was validated by the determination coefficient
(R2) and adjusted R2 is 0.9674 and
0.9349 respectively. In this case, the value of R2(96.7%) indicated that the sample variation of 96.7% for pH was
attributed to the independent variables and that 3.3% of the total
variation could not be explained by the model. The R2and the adjusted R2 for pH are highly adequate because
they have satisfactory levels of R2 of more than 80%
with no significant Lack of Fit [11, 13]. The model shows that TTA
had a coefficient of determination (R2) of 0.7846 and
the adjusted R2 of 0.5692. which means that 78.46% of
the variability in the response could be explain by the model, This
might be due to experimental errors, handling errors etc.