A semi-automatic system for classifying and describing foods according to FoodEX2
Authors
T. Eftimov, G. Ispirova, P. Korošec, B. Koroušić Seljak
Publication
Metrology Promoting Harmonization & Standardization in Food & Nutrition 3rdIMEKOFOODS
, Thessaloniki, Greece, 1-4 October, 2017
Abstract
In this paper, we present results of the evaluation of a semi-automatic system for classifying and describing foods according to FoodEx2 using datasets from three European countries. The proposed system is an integration of methods from machine learning, natural language processing and probability theory. It obtained an accuracy of 91% for the classification part and 78% for the description part. The usage of the system can be a link between food consumption and food composition data as the transformation from food intake into nutrient intake can be automatically made.
BIBTEX copied to Clipboard