The HOC-TV project intends to apply heuristic optimisation techniques in order to find the best clustering of TV audience preferences: Given raw data about past programming grids and viewers’ program preferences, data mining (or KDD) methods are applied to identify useful and understandable features. These features are then used to form groups (clusters) in such a way that the individuals inside the same group have the most similar TV programming preferences, in other words, will choose to watch the same TV channel. Clustering algorithms are used to obtain partitions of the TV preferences.
Started in January 2000 and was concluded in 2000.
Participating entities: CENTRIA - UNL, Univ. Évora.
Funding entity: PRAXIS.
Principal researcher: Fernando Moura Pires.
2 papers during 2000, and implementation of a prototype.
