Recently, several types of artificial neural networks (ANNs) have been applied to classify a specific dataset. However, most of these models use only a limited number of features as input, in which case there may not be enough information to make the prediction due to the complexity of the starting dataset. If you have more features, the run time of training would be increased and generalization performance would deteriorate due to the curse of dimesionality. In these cases, a tool to extract the characteristics would be particularly useful. RBM is a machine learning tool with a strong representation power, which is often used as a feature extractor in a wide variety of classification problems.