The global adoption of information technology systems throughout all activity domains lead to the existence of a huge amount of information systems unable to interoperate, first of all, because of different ways of identifying resources. Different transport and application level protocols for data exchange also raise some interoperability difficulties. Common interoperability scenarios rely on tightly controlled, specific communities of information technology islands. In this chapter we present and discuss the usage of lexical, syntactic and semantic lexical technologies to address interoperability problems at the product identification level, in the context of food consumption analysis.
TopBackground
The possibility of developing a computer application that allows automated alignment between food consumption/sales databases and food composition databases that contain their nutritional composition or other properties (e.g. contaminants, allergens) is seem of extreme value by public health care authorities.
Many efforts have been done in developing food classification and description systems, as well as the corresponding creation of controlled vocabularies, in order to allow food consumption analysis and related applications.
Although this represents a recurring issue/problem in this scientific area, no system or software application is known for enabling the automatic mapping of food descriptions between sales and food composition databases.
All known applications addressing these issues require the usage of pre-defined identifiers, pre-defined taxonomies and specific, controlled vocabularies for products identification, classification and description. One example of this type of approach is the FoodEx2 (European Food Safety Authority, 2011) specification done by the European Food Safety Authority. FoodEX2 aims to support food consumption, electronic transmission of chemical occurrence data and public health risk assessment in the framework of public health policies for the European Union countries.
In the context of our study, this type of initiatives is seen as feasible only for a small scale number of participants that are able to cope with the technical alignment and cost of adapting their systems.
Other type of approaches aiming for semantic integration of food related information are also worth to mention. Among those approaches, the semantic search engines and the semantic food repository systems are the most relevant in the context of our study.
The search engines could be highly enhanced with the aid of semantic tools. After that evolution we can expect that the search engines could be more intelligent, being able to understand, not only some predefined keywords, but also complex human-readable sentences.
The Knigine, the Hakia Search Engine and the DuckDuckGo are examples of search engines which already take advantage of Web semantic technologies. As in our work, this allows them to provide a broader and a more intelligent set of results.