Research on Model for Sensory Quality of Yogurt Based on Bagging

Abstract

Yogurt is a common dairy product in daily life. How to quickly and accurately identify the sensory quality of yogurt is of great significance to the control of sensory quality of yogurt. In this paper, sensor data of 120 yogurt samples were obtained by electronic nose, and the measured sensor data were used as inputs to construct 2-layer back propagation neural network(BPNN) models. Then the Bagging method was employed to integrate the BPNN models, which constructed the sensory quality classification model for yogurt. The comparative experiment showed that the sensory quality classification model based on Bagging-BPNN has better accuracy and generalizability than the model based on single BPNN and k-nearest neighbors(kNN) algorithm.