Applying Feed-Forward Neural Networks to Collaborative Filtering

Neural networks have been applied to an entire plethora of learning tasks, such as text recognition, credit rating analysis and prediction, and so forth. Surprisingly, there have never been proposals suggesting their usage for collaborative filtering (CF) scenarios. The objective of this diploma/master’s thesis is to close that gap by virtue of (i) conceiving and applying neural network models to suit CF problems and (ii) evaluating their fitness with respect to benchmark approaches, taking into the effect of diverse tuning parameters.