We tried to understand the differences between true-italian and pseudo-italian recipes.
Based on expert knowledge of which recipe is true-italian and which is pseudo-italian we tried to find the italian-factor.
Prototype pseudo-italian recipes include Spaghetti Napoli, Penne Alfredo, Meatballs-Pasta and Pizza Hawaii.
The analyses are based on a tidy list of selected recipes, their ingredients and amounts.
The average ingredient-amount of all ingredients which are present in at least one true- and one pseudo-italian recipe were compared:
There is no clear italian-factor apparent.
But small differences were visible. Amongst others, Pseudo-italian recipes seem to contain a bit more beef, more tomatoes and a bit less spaghetti.
Using PCA, we hoped to find two clusters containing the true- and pseudo-italian recipes.
There is a cluster of true-italian recipes.
But the problem is highly complex, as the eigenvalues/variances of dimensions show:
A clear italian-factor still remains a mystery. We failed to find a defining factor with the applied methods.
If one would find the italian-factor, which defines a true-italian recipe, one could make predictions for any unknown recipe based on the ingredients/amounts.
We thank ozagordi for his continuous support and advice in conducting this study.
sschmutz, vkufne and mihuber collected the data; sschmutz analyzed the data and wrote the manuscript; vkufne and mihuber reviewed and approved the manuscript.