Progress in Research on Machine Learning for Studies on Food Flavor Analysis

Abstract

The flavor profile is an important factor that affects consumer preference and purchase intention and plays a vital role in food production as well as in food science research. Traditional methods for determining food flavor are based on sensory evaluation, instrumental analysis, or a combination of the two. With the rapid development of computer technology, machine learning technology with high predictive ability and accuracy has been widely used for food flavor analysis and prediction. This method overcomes the limitations of traditional methods in food flavor evaluation, which are time-consuming and cannot be used to process large quantities of data. This review discusses the latest progress in food flavor analysis technology and machine learning technology in food flavor research and presents commonly used food flavor analysis techniques and machine learning algorithms. Based on an overview of common machine learning models, this review systematically summarizes the application of machine learning in the high-throughput screening of food flavor substances, flavor perception, and flavor quality control. The challenges and future research trends in the application of machine learning to food flavor analysis and prediction are summarized and discussed. The purpose of this review is to provide theoretical references for machine learning technology to predict flavor formation mechanisms, synthesize desirable flavor compounds, and control food quality.

Keywords

• food flavor analysis
• machine learning
• high-throughput screening
• flavor prediction
• quality control