[4J1-OS-25a] 食とAI(1/2)

[4J1-OS-25a-03] Modeling Texture Topics from a Recipe Sharing Site Combined by Quantitative Characteristics.

〇Hiroshi Uehara¹, Daichi Mochihashi² (1. Rissho University, 2. The Institute of Statistical Mathematics)

Keywords:Recipe sharing site, Rheology of gels, Topic model, Texture expressoin, Texture profile analysis

This study tries to estimate texture of gel related dishes based on both the data of a recipe sharing site, and research results of food science. Most of recipes are accompanied by insufficient information about what kind of texture they realize. We propose a method to estimate characteristics of texture for each recipe by applying a joint topic model to bridge sensory texture terms in a recipe sharing site and corresponding quantitative texture resulting from food science research. The result shows estimated texture terms for dishes are coincident with rheology, the quantitative texture provided by related food science research.