Deep Autoencoder for Recommender Systems: Parameter Influence Analysis

Recommender systems have recently attracted many researchers in the deep learning community. The state-of-the-art deep neural network models used in recommender systems are multilayer perceptron and deep autoencoder (DAE). In this work, we focus on the DAE model due to its superior capability to reconstruct the inputs, which works well for recommender systems. Existing works have similar implementations of DAE but the parameter settings are vastly different for similar datasets. In this work, we have built a flexible DAE model, named FlexEncoder that uses configurable parameters and unique features to analyze the parameter influences on the prediction accuracy of recommendations. Extensive evaluation on the MovieLens datasets are conducted, which drives our conclusions on the influences of DAE parameters. We find that DAE parameters strongly affect the prediction accuracy of the recommender systems, and the effect remains valid for bigger datasets in the same family