Temporal rating habits
Pedro G. Campos, Fernando Díez, Alejandro Bellogín · 2011
In this paper, we describe the experiments conducted by the Information Retrieval Group at the Universidad Autónoma de Madrid (Spain) to tackle the Identifying Ratings (track 2) task of the CAMRa 2011 Challenge. The experiments performed include time-frequency probabilistic strategies, heuristic collaborative filtering (CF) and a model-based CF approach. Results show that probabilistic classifiers based on temporal behavior of users have better performance than traditional recommendation-based strategies, thus reflecting that temporal information is a valuable source for the identification or discrimination of user ratings.