MULTIMEDIA DISTRIBUTION SYSTEM WITH USER BEHAVIOR ANALYZER

A multimedia distribution system or interactive digital television broadcast system (iDTV) comprising several content providers (CP: CP1, CP2) broadcasting channels of video signals to a user (TV, STB) via a content selector (CS). The broadcast system comprises a user behavior analyzer (UBA) with quantitative behavior classifier (E(w)) able to analyze and to classify the quantitative consumption behavior of the user, and with a qualitative behavior classifier (F(x)) having as output classified user behaviors and prediction on future behaviors of the user. The classifiers are coupled to an opportunities selector (G(y)) that also receives information from additional content providers (ACP: ACP1, ACP2, ACP3). Under control of the opportunities selector G(y), additional content, e.g. advertisements, meta-information, iTV applications, trailers, ... is send to a multimedia data combiner (H(z)) whereby the additional content is published, e.g. by mixing or by an overlay operation on/ over the selected channel. As a result, the end-user gets his broadcast channel adapted with the content that was selected based on his consumption behavior. The role of the qualitative behavior classifier classifying and predicting the user consumption behaviors is different from the role of the opportunities selector defining, predicting and classifying windows of opportunities to publish/ add content from additional content providers; and the interface (I2) between them has an important role as business-to-business (B2B) interface. Cette invention concerne un système de diffusion multimédia ou un système de télévision numérique interactive (iDTV) comprenant plusieurs fournisseurs de contenus (CP: CP1, CP2) diffusant des canaux de signaux vidéo à un utilisateur (TV, STB) par l'intermédiaire d'un sélecteur de contenu (CS). Le système de diffusion comprend un dispositif permettant d'analyser le comportement d'un utilisateur (UBA) pourvu d'un dispositif de classification quantitative du comportement (E(w)) pouvant analyser et classifier l'importance du comportement de consommation de l'utilisateur, et d'un dispositif de classification qualitative du comportement (F(x)) conçu pour fournir des comportements utilisateur classifiés et pour prédire des comportements utilisateur futurs. Les dispositifs de classification sont couplés à un sélecteur de possibilités (G(y)) qui reçoit également des informations transmises par d'autres fournisseurs de contenus (ACP: ACP1, ACP2, ACP3). Sous le c