Data on physicochemical analysis (pH, total acidity, and the contents of total soluble solids, residual sugar, ethanol and volatile flavor compounds) and microbiological enumeration (fungi, yeasts and bacteria) of the resulting soy sauce were statistically analyzed using one-way ANOVA procedure with SPSS ver. 16.0 software (SPSS, Chicago, IL, USA). Significant differences were determined using Duncan's multiple range test at the level of p < 0.05. Principal component analysis (PCA) with varimax rotation was used to reduce the dimensionality of the data set to a small number of PCs with high correlation. PCA bi-plots of major principal components were drawn to demonstrate the differences and similarities among the samples. Hierarchical cluster analysis (HCA) with Euclidean
distance and average between-groups linkage was also used to
classify the fermented soy sauce samples into subgroups that had
similar physicochemical properties.