Computational gastronomy: the mathematics of cooking

Abstract

We propose to build a framework for the generative grammar of cooking analogous to that of languages (Bagler, arXiv:2211.09059, 2022). By defining a cooking recipe as a finite state language, we intend to map elements of culinary instructions to English language sentences and express cooking as a Markov model. Taking the probabilities a step further, we calculated the support vectors for the ingredients from the recipe data. This research project explores Indian cuisine by analysing recipe similarities through ingredient-based data. By examining the ingredients used in various Indian recipes and applying similarity calculations, the study categorises recipes into clusters based on their likeness. The aim is to uncover recurring patterns regional variations, and develop a recommendation system for exploring similar recipes. This analysis provides insights into Indian culinary traditions and has implications for recommendation systems and cultural studies.