Beef sensory qualities prediction based on stepwise regression method

The aim of this study was to analyze the correlation between sensory qualities and physiochemical properties of beef,and to establish validate model for predict sensory quality.Longissimus doris,rectus femoralis,biceps femoralis,vastus lateralis,vastus intermedius and tritip muscle from 9 cows( total 54 samples) were selected and consumer sensory evaluation and physiochemical properties were measured. Linear relationship between physiochemical properties and sensory evaluation were analyzed. The results showed that most of the physiochemical properties were significantly or marked significantly correlated( p < 0.05 or p < 0.01) with sensory evaluation.Stepwise regression analysis was used to generate prediction equations with the data of physiochemical traits as independent variables respectively and the data of sensory analysis as dependent variable.The prediction equations of sensory tenderness,juiciness,flavor and overall liking were of significance in statistics and the determination coefficients were 0.55,0.37,0.68 and 0.36,respectively. The sensory qualities of beef could be well described by the validate model.