Prediction of Sensory Evaluation Values of Cookies Using Machine Learning

User Benefits

━ Texture testing of food products is possible by using the EZTest texture analyzer. ━ Various jigs can be exchanged easily by using the jig platform. ━ It is possible to predict approximate sensory evaluation values by using machine learning.

Introduction

The factors that humans experience as the “deliciousness” of food include factors related the food itself (e.g., flavor, aroma, food texture) and human factors (physiological and psychological factors, eating habits, external factors), but because texture accounts for a large part of the perception of deliciousness, depending on the food, the evaluation of food texture has become one important item in food testing. Methods for evaluating food texture can be divided into sensory evaluation, in which human subjects evaluate the “mouthfeel” of the food when eaten, and evaluation of physical (mechanical) properties such as the hardness of the food using instruments. Mouthfeel is generally evaluated by sensory tests, but the difficulty of reproducing the evaluation results due to individual differencesin the human senses and the physical condition of the subjects is an issue in sensory tests. For this reason, measurements using instrument are conducted in order to obtain objective results. As representative mechanical properties of texture, Fig. 1 shows texture profile proposed by Szczesniak. Although the texture profile makes it possible to evaluate the basic mouthfeel of foods, it is difficult to measure more complex characteristics of mouthfeel.

January 16, 2024 GMT