- Volume 16 Issue 2
Flavor components extracted from eighty species of Korean ordinary soy sauce were analyzed by gas chromatography. The relationship between the sensory scores of soy sauce flavor and the gas chromatographic data transformed with variables were analysed by method of multiple regression analysis. Simple correlation between values of each peak and sensory scores were totally low. The tenth and 12th peak had the highest correlation, 0.331. Determination coefficients of data obtained by transformation of each variables were not significantly different from each other. Flavor of soy sauce was explained about 56% at step 16 in case of stepwise multiple regression analysis of absolute values. The fact that the minimum standard errors of an estimate was found at the 16th step suggests the importance of selecting of independent variables from the whole gas chromatogram together with the results of F ratio. In the contributing proportion of each peak examined, peak 10 and 12 were contributing mainly to the good flavor of soy sauce.