情報処理学会 第77回全国大会講演要旨

Food evolution had been one of the most telling witnesses of the human beings' living activities. Knowing food history helps us predict the future of food preferences. Before, an individual researcher could track changes of food by manual investigation of recipes, since propagation of information among common people was slow, usually limited to be narrow and regional. Today, common people are able to upload their own recipes to recipe sites like Cookpad in Japan easily. The information of food has been increased rapidly and tracking food evolution becomes impossible without computer-based data analysis. To help understanding the evolution of food preference, we propose a novel measure, called traditionality score, of a recipe, and propose a method for computing the score by analyzing recipe data including characteristics of the ingredients of each recipe. Employing recipe data of several kinds of foods on Cookpad, we conduct case studies through experiments of computing rankings according to the traditionality score.