Complexity and Similarity of Recipes based on Entropy Measurement

Abstract

Background/Objective: New cuisines are being constantly created by people and shared on the internet. Cooking is usually classified and evaluated by subjective factors such as the personal cooking ability. Methods/Statistical Analysis: In this paper, we define the objective method for measuring the recipe complexity using the probabilistic entropy measurement. Through the cooking ingredients and cooking verbs in a recipe, we measure the recipe complexity in terms of the recipe preparation and the recipe procedure. And we calculate the recipe similarity by the entropy of common ingredients and construct the social network of recipes that observes the whole correlation between recipes. Findings: As a result, the most difficult recipe in the cooking preparation is ‘Braised Dongtae Seafood’ that needs 27 ingredients. And the easiest recipe is ‘Boiled Potato’ that only needs 3 ingredients: ‘Potato’, Salt’ and ‘Sugar’. In the cooking procedure, ‘Jeolla-Province Mosi-Songpyeon’ is the most difficult recipe that uses 18 verb types and 28 verbs. The easiest recipe is ‘Healthy Silk-Fowl Soup’ that only uses ‘Boil’ to cook. Through the recipe network, we can measure the distance of the recipes and minimize the effort in ingredient preparation and cooking procedure when we planed the cooking schedule to prepare for many dishes. Application/Improvements: If we plan to cook a schedule, the easy path to cook could be indicated by the shortest path between the recipes in the network.