Inferring Food Preferences and Dietary Structures of the Americans with Transfer Learning

Yihan Chen, Lin Jiang, Zhen Tan

Research output: Chapter in Book/Conference proceedingBook Chapterpeer-review


Health problems have always received significant attention. According to previous research, different food types are highly related to multiple chronic diseases. However, few studies have been conducted to comprehensively assess the future development of people’s health in terms of food structure or food preferences. This paper uses transfer learning-based image recognition method to analyze the food preferences of the United States people for 101 kinds of food based on online rating information. According to the analysis results, the overall food preferences of the Americans tend to be healthy, although foods with high sugar content are still one of the top-rated categories among the public. In addition, compared to domestic foods, foreign foods are widely popular among American residents. Our study can be useful for the research perspective that food preference reflects people’s tendency of future food structure and, based on which, predicts their future health.
Original languageEnglish
Title of host publicationTransfer Learning - Leveraging the Capability of Pre-trained Models Across Different Domains
EditorsAnwar P.P. Abdul Majeed
Publication statusPublished - 14 May 2024


  • food preference
  • dietary structure
  • star ratings
  • image classification
  • transfer learning


Dive into the research topics of 'Inferring Food Preferences and Dietary Structures of the Americans with Transfer Learning'. Together they form a unique fingerprint.

Cite this