2.4. Statistical analysis
All the results were expressed as mean of six parallel readings along
with standard error. Statistical significance of the data set was adjudged
in SPSS version-20 software (IBM SPSS Statistics, USA) at 95% confi-
dence interval. Quadratic polynomial model fit was performed through
nonlinear regression in Microsoft Excel 2010 (Microsoft Office system,
USA). Analysis of variance (ANOVA) of the data set was also performed
and the model accuracy was expressed by computing correlation coeffi-
cient (R2
) and adjusted R2
.
Multivariate analysis of variance (MANOVA) was performed for six
different responses (TCD, BI, TPC, TAC, AA loss and TFC loss) affected
through processing in SPSS-20. Principal component analysis (PCA)
was applied for the data corresponding to those six responses; the number
of factors were reduced to a level by which maximum variability can
be explained in the data set. Relativity within the data set of different
factors was also determined through correlation matrix.