Personalized Mass Media


Today's media companies are confronted with HyperConnected visitor, readers and subscribers. People use their IP-TV, PC, iPads and smartphones to look at content, read news, look at multimedia content... . Each interaction with a visitor contains information about what the user's interests are, where he is, what he likes to read or view depending on devices, time of the day and even location. Current systems fail in capturing all that information and create some insight in the vast amount of data. This project targeted to obtain that insight and use it in an actionable way.


Given the real-time aspect and the amount of content and interactions we used a big data system combined with machine learning techniques like collaborative filtering in order to provide a personalized content experience. Combined with content enhancement techniques like named entity recognition we  lift the quality of the recommendations. On top of that we provide a real-time dashboard where the news editor can track user activity on the content, recommendations and users.


Year: 2012 - 2013
Partners: VRT, NGDATA, LimeCraft
Technologies used: NGDATA's Lily Enterprise, Named Entity Recognition Apache Stanbol, Atex Polopoly


Innacco Innovation and Enterprise Architecture Services

Contact us

innacco bvba
Steilvoordehof 12
9070 Heusden