LanguaL Food Description: a Learning Process

Abstract

The European Food Information Resource (EuroFIR) Network of Excellence (http://www.eurofir.eu) aims to provide validated food composition data (FCD) from European food composition databanks. However, the network covers 27 countries with different languages and food cultures, making comparisons difficult. Moreover, in 2005, only 25% of the databases included international food classification or food description. To overcome some of the challenges, it was decided to use the LanguaL thesaurus (http://www.langual.org) to index (systematically describe) and link the foods in European food composition databases (FCDBs). To facilitate food indexing, LanguaL Food Product Indexer software was introduced and several short (1-2 days) food indexing courses for FCD compilers from all participating countries were organised. Feedback between the LanguaL Technical Committee and the FCD compilers allowed the latter to improve their food indexing skills. In turn, the compilers proposed new descriptors and translations for the thesaurus. The result was a set of more than 26,000 foods in national databases and 2360 foods in specialised data sets that were LanguaL indexed and thus able to be linked to the EuroFIR network. Both the EuroFIR network and the individual FCD compilers benefit from standardised food description, allowing foods to be linked and compared across borders and language barriers. The LanguaL thesaurus has, in turn, benefited from the expertise of the FCDB compilers from different cultures.