Description

The R&D result supplier has implemented semantic technologies for automatic analysis of textual content consumed by the user of an internet services platform (news articles, ads, tags, metadata in rich media etc) and consequently the production of personalized recommendations based on lightweight ontological knowledge structures. For example, regarding a user that consumes content associated to “politics” and “travel”, following the automatic modeling of his/her profile, the system can provide personalized ranked recommendations on related products/services/content. The methodology encompasses a set of distributed technologies for text extraction and analysis in order to semantically describe the domain in question, as well as for automated extraction and semantic classification of user preferences and finally for recommending content to the user, including:

-          Advanced analysis of the semantic relations between content-extracted terms/tags/metadata.

-          Semantic classification of raw textual content based on predefined ontological knowledge bases

-          Automatic and unobtrusive extraction of user web behavior based on his/her interaction to a digital services platform

-          Semantic modeling/learning of detected user preferences

-          Ranked content recommendation via semantic matching

 The advantages and main innovations offered by the proposed technologies include:

-          Automatic, more complete and meaningful description of data

-          High performance on analyzing and classifying content

-          High performance on efficiently retrieving offered products/services

-          Lightweight and uniform semantic structures minimizing load on mass data trafficking and allowing for:

-          Profiling and recommendation on the client side, thus protecting user privacy and alleviating server communication workload

-          Easy parameterization across different domains (leisure, news etc)

 

·   Identification of R&D activity

Acronym: Semantic Personalised Recommender
Author of the R&D result:  Dr. Ioannis Kompatsiaris, Dorothea Tsatsou
Company or University: Centre for Research and Technology Hellas, Informatics & Telematics Institute, Multimedia Group
Phone: Tel: +30 2311 257774 (I. Kompatsiaris)
E-mail:  ikom@iti.gr
Sector:  ICT
Type of the R&D result:  Software platform

View a detailed description in the R&D Repository »

Request a proposal

Valorisation Plan Authors

Related Documents

There in no related documents

Visit the other applications of the INTERVALUE Platform: R&D Repository | IP Agreements

© 2009-2010 INTERVALUE Project