Collaborative ensemble learning: combining collaborative and content-based information filtering via hierarchical bayes

摘要: Collaborative filtering (CF) and content-based (CBF) have widely been used information applications, both approaches having their individual strengths weaknesses. This paper proposes a novel probabilistic framework to unify CF CBF, named collaborative ensemble learning. Based on content based models for each user's preferences (the CBF idea), it combines society of users' predict an active idea). While retaining intuitive explanation, the combination scheme can be interpreted as hierarchical Bayesian approach in which common prior distribution is learned from related experiments. It does not require global training stage thus incrementally incorporate new data. We report results two data sets, neuters-21578 text set base user opionions art images. For achieved excellent performance terms recommendation accuracy. In addition engines, learning applicable problems typically solved via classical Bayes, like multisensor fusion multitask