1,683
edits
Changes
→Some ideas to not for later and maybe add to this page:: new section
Because it strongly overlaps with a theoretical "diet tracking" tool
Because it strongly overlaps with a theoretical "diet tracking" tool
== Some ideas to not for later and maybe add to this page: ==
From https://www.researchgate.net/publication/282973643_Insights_from_Machine-Learned_Diet_Success_Prediction.
"The lack of con-
Pacific Symposium on Biocomputing 2016
543
trolled vocabulary in the food data creation causes several data integrity issues, e.g., the same
foods may be described slightly different by different users. To mitigate the problems, we
needed ways to group related foods into semantic categories. Thus, we built a food taxonomy
by compiling lists of food-related categories and page names from Wikipediac. The taxonomy
is manually organized into 18 main categories (e.g., staple food, meats, vegetables, etc.), 149
subcategories (e.g., wheat, rice, beef, etc.) and 4,233 entities, describing individual ingredi-
ents and meal types of food entries. For example, the entry “McDonald’s - Premium Sweet
Chili Chicken Wrap (Grilled)” will be annotated with the following set of {main category:
subcategory: entity}:{Staple foods: Wheat: Wrap},{Meats: Poultry: Chicken},{Preparation
Methods: Grill},{Fast foods: McDonald’s}."