Evaluation and Assessment of Recommenders Using Monte Carlo Simulation

There have been various definitions, representations and derivations of trust in the context of recommender systems. This article presents a recommender predictive model based on collaborative filtering techniques that incorporate a fuzzy-driven quantifier, which includes two upmost relevant social phenomena parameters to address the vagueness inherent in the assessment of trust in social networks relationships. An experimental evaluation procedure utilizing a case study is conducted to analyze the overall predictive accuracy. These results show that the proposed methodology improves the performance of classical recommender approaches. Possible extensions are then outlined