COMPSTAT 2018: Start Registration
View Submission - COMPSTAT2018
A0403
Title: Exploratory data analysis about the optimal number of hidden layers and nodes in deep neural network Authors:  Jae Eun Lee - Pukyong National University (Korea, South) [presenting]
Dae-Heung Jang - Pukyong National University (Korea, South)
Abstract: To find out the optimal network configuration about the number of hidden layers and nodes in deep neural network, we need to use the exploratory data analysis (EDA) approach. This is because it is difficult to discover theoretical approach. We perform EDA for finding the optimal network configuration about the number of hidden layers and nodes in deep neural network through some examples of the prediction and the classification. When we try the prediction and the classification about datasets using deep neural network, we examine whether the deep neural network can decide the optimal number of hidden layers and nodes through the EDA approach.