Abstract: Food flavor is a key factor influencing consumer choice and satisfaction. Traditional methods for food flavor analysis suffer from limitations such as low efficiency and strong subjectivity. In contrast, machine learning technology, with its powerful data processing and pattern recognition capabilities, offers new solutions for food flavor research. This paper reviews machine learning algorithms employed in food flavor research and flavor analysis strategies incorporating machine learning. It outlines commonly used analytical instruments and explores flavor analysis methodologies at three distinct levels: molecular structure, properties, and sequence. Ultimately, a reliable and comprehensive flavor analysis pathway has been established using existing flavor databases as methodological supplements. The paper also proposes improvements to address the shortcomings of machine learning methodologies, providing directional references for food flavor analysis strategies. This contributes to advancing the transformation and upgrading of the food industry.

Key words: food flavor analysis strategies; machine learning; molecular structure; molecular properties; molecular sequence; deep learning; flavor molecules