A topological analysis of the space of recipes

Analyzing Culinary Recipes Using Topological Data Analysis

Title: A topological analysis of the space of recipes

Authors: Emerson G. Escolar, Yuta Shimada, Masahiro Yuasa

Source: Kobe University Graduate School of Human Development and Environment, Japan

Introduction

The paper investigates the use of topological data analysis (TDA), particularly persistent homology, to decipher the structure of the space of culinary recipes. Culinary recipes have become a subject of computational analysis aimed at understanding underlying patterns and generating novel ingredient combinations. Traditional network science methods have been useful, but this study introduces a topological perspective for deeper insights.

Methodology

Persistent homology, a key tool in TDA, is utilized to identify multiscale features such as connected components and "holes" in the space of recipes. The data consists primarily of combinations of ingredients from the Supplementary Dataset 2 of Ahn et al. (2011). The cosine dissimilarity measure was chosen to quantify the relationships between recipes. This measure takes into account the angular dissimilarity between recipes represented as 0-1 vectors.

The space of recipes is treated as a Vietoris-Rips complex, where simplices are formed based on the dissimilarities between recipes. Persistent homology then tracks the birth and death of topological features across different thresholds, resulting in a persistence diagram which provides a summary of the topological structure of the data.

Key Contributions

1. Topology in Recipe Analysis: The study introduces the application of TDA to culinary recipes, offering a novel way to explore the structural characteristics of recipe space.
2. Cycle Identification: TDA identifies cycles in the data that correspond to multiscale "holes," revealing gaps in the recipe space that could potentially be filled with new combinations.
3. Novel Recipe Generation: By exploiting the identified cycles, the authors propose new ingredient combinations using combinatorial optimization, focusing on maximizing the dissimilarity to existing recipes while maintaining coherence based on topological information.

Results

Persistent homology results are displayed in persistence diagrams, which show several significant birth-death pairs outliving others, indicating the presence of prominent topological features. The authors focus on the top 5% of birth-death pairs with the longest lifespans for further analysis.

From these identified cycles, the authors generate novel ingredient combinations. For example, they identified combinations such as cranberry, cream cheese, gin, raisin, and whole grain wheat flour as potential new recipes. Several of these combinations were used to create different variants of cream cheese biscuits.

Practical Implications

A sensory evaluation was conducted to validate the acceptability of the new recipes, confirming that the novel combinations were palatable. This practical outcome demonstrates that topological insights can be translated into viable culinary innovations.

Theoretical Implications

The work establishes a new paradigm for analyzing culinary recipes, moving beyond traditional network-based methods to embrace a topological framework. This approach not only identifies gaps in the data but also provides a principled method for generating novel combinations.

Future Developments

The authors suggest further refinement of the methods by incorporating more detailed aspects of recipes, such as cooking methods and ingredient quantities. Additionally, integrating flavor profiles into the topological analysis could provide a more robust framework for computational gastronomy.

Conclusion

This exploratory study highlights the potential of TDA in computational gastronomy. By providing new tools to understand and navigate the space of culinary recipes, it paves the way for innovative recipe generation. Future research in this direction could significantly enhance the repertoire of creative chefs, aiding in the emergence of novel and delightful culinary creations.