Ontology-based Recommender System of Economic Articles
This addresses the need for personalized economic information for decision-makers at a specific company, representing an incremental improvement in domain-specific recommendation systems.
The paper tackles the problem of information overload in economic press reviews by proposing an ontology-based recommender system that semantically matches articles with user profiles, aiming to provide customized reviews for clients.
Decision makers need economical information to drive their decisions. The Company Actualis SARL is specialized in the production and distribution of a press review about French regional economic actors. This economic review represents for a client a prospecting tool on partners and competitors. To reduce the overload of useless information, the company is moving towards a customized review for each customer. Three issues appear to achieve this goal. First, how to identify the elements in the text in order to extract objects that match with the recommendation's criteria presented? Second, How to define the structure of these objects, relationships and articles in order to provide a source of knowledge usable by the extraction process to produce new knowledge from articles? The latter issue is the feedback on customer experience to identify the quality of distributed information in real-time and to improve the relevance of the recommendations. This paper presents a new type of recommendation based on the semantic description of both articles and user profile.