ECML PKDD Discovery Challenge Recommending Given Names

Recommender systems are data filtering systems that suggest data items of interest by predicting user preferences. In this paper, we describe the recommender system developed by the team named uefs.br for the offline competition of the 15th ECML PKDD Discovery Challenge 2013 on building a recommendation system for given names. The proposed system is a hybrid recommender system that applied a content-based approach, a collaborative filtering approach and a popularity approach. The final recommendation is composed by the results of these three different approaches, in which parameters where optimized according to experiments conducted in datasets built from train data.