Deep learning based recommender systems

In parallel with the rapid development of prospective systems in the last 20 years, many methods have been applied to this field. One of them is the deep learning networks that have attracted the interest of researchers in recent years. The DBN (Deep Belief Network), which trains one layer at a time greedily, uses unsupervised learning for each layer and is composed of RBMs (Restricted Boltzman Machine), has become a turning point in this area. In this study, the deep learning method is applied to the recommender system problem. The Python-based deep learning library, Keras, is used and the existing learning algorithms are compared.