Decoding Global Palates: Flavor Preferences in Online Recipes

This academic article explores cross-cultural flavor preferences by analyzing a large dataset of online recipes from China, the US, and Germany. The authors utilize flavor compounds extracted from ingredients to represent recipes and employ machine learning to identify distinctions and similarities in flavor use across these cultures. Their findings suggest cultural differences in flavor profiles are detectable algorithmically, and they also uncover potential shared flavor preferences between some cultures. The research further includes exploratory analyses using ingredient networks and flavor compound clustering to support the machine learning results and discusses implications for cross-cultural food recommendation systems