Learning Rate Estimation Model in Restricted Boltizmann Machine Neural Network for Building Recommender Systems

Abstract

This paper finds the mechanism to obtain learning rate value for training restricted Boltzmann artificial neural network that used for build recommender systems. One of the important problem in training the artificial neural network model is finding the appropriate learning rate values for making designed model reach the optimal result for learning process. The proposed model analyzes the behavior of the recommender system in context of mean squared error (MSE), mean absolute error (MAE), precision and recall as well as the free energy function. Proposed model