Ingredient-flavor Network

In network science, ingredient-flavor networks are networks describing the sharing of flavor compounds of culinary ingredients. In the bipartite form, an ingredient-flavor network consist of two different types of nodes: the ingredients used in the recipes and the flavor compounds that contributes to the flavor of each ingredients. The links connecting different types of nodes are undirected, represent certain compound occur in each ingredients. The ingredient-flavor network can also be projected in the ingredient or compound space where nodes are ingredients or compounds, links represents the sharing of the same compounds to different ingredients or the coexistence in the same ingredient of different compounds.

History

In 2011, Yong-Yeol Ahn, Sebastian E. Ahnert, James P. Bagrow and Albert-László Barabási"Flavor network and the principles of food pairing" by Yong-Yeol Ahn, Sebastian E. Ahnert, James P. Bagrow & Albert-Laszlo Barabasi in NATURE, SCIENTIFIC REPORTS 1 : 196, DOI: 10.1038/srep00196
investigated the ingredient-flavor networks of North American, Latin American, Western European, Southern European and East Asian cuisines. Based on culinary repository epicurious.com,epicurious.com
/ref> allrecipes.comallrecipes.com
/ref> and menupan.com,menupan.com
/ref> 56,498 recipes were included in the survey.

The efforts to apply network analysis on foods also occurred in the work of Kinouchi"The non-equilibrium nature of culinary evolution" by Osame Kinouchi, Rosa W Diez-Garcia, Adriano J Holanda, Pedro Zambianchi & Antonio C Roque in New Journal of Physics 10 (2008), 073020.
/ref> and Chun-Yuen Teng,"Recipe recommendation using ingredient networks" by Chun-Yuen Teng, Yu-Ru Lin & Lada A. Adamic in Proc. 3rd Annu. ACM Web Sci. Conf.(WebSci’12), Jun. 2012, pp. 298–307.
/ref> with the former examined the relationship between ingredients and recipes, and the latter derived the ingredient-ingredient networks of both compliments and substitutions. Yet Ahn's
ingredient-flavor network was constructed based on the molecular level understanding of culinary networks and received wide attentionMIT Technology Review
/ref>Wired.com
/ref>

Properties

According to Ahn, in the total number of 56,498 recipes studied, 381 ingredients and 1021 flavor compounds were identified. On average, each ingredient connected to 51 flavor compounds.

It was found that in comparison with random pairing of ingredients and flavor compounds, North American cuisines tend to share more compounds while East Asian cuisines tend to share fewer compounds. It was also shown that this tendency was mostly generated by the frequently used ingredients in each cuisines.

Food pairing

An important feature that the ingredient-flavor network showed is the principle of food pairing. A well known hypothesis states that ingredients sharing flavor compounds are more likely to taste well together than ingredients that do not. However, the sensory test by Miriam Kort, etc. claimed that the shared compound hypothesis can be debatable.
According to Ahn, the food pairing pattern changes in different cuisines. North American recipes tends to obey the shared compound hypothesis while East Asian cuisines tend to avoid it. Besides the spatial variance, Kush R. Varshney, Lav R. Varshney, Jun Wang, and Daniel Myers "Flavor Pairing in Medieval European Cuisine: A Study in Cooking with Dirty Data" by Kush R. Varshney, Lav R. Varshney, Jun Wang, and Daniel Myers in Proc. Int. Joint Conf. Artif. Intell. Workshops, Aug. 2013, pp. 3–12
/ref> also showed the time variance in food pairing by comparing the modern European recipes with the Medieval European recipes. They concluded that the Medieval cuisine tend to share more compounds than the cuisine today.

See also

* Albert-László Barabási
* Bipartite graph
* Bipartite network projection
* Food science
* Graph theory
* Network science
* Network theory
* Sensory analysis

References

{{DEFAULTSORT:Ingredient-flavor networks
Application-specific graphs
Flavors
Food science