Prediction of Manufacturing Plant's Electric Power Using Machine Learning

In this paper, we apply the data accumulated through E-IOT platform to machine learning method to find significant variables first and predict the electric power generated in manufacturing process by using these variables. Pre-processing such as resampling of data was carried out before the prediction. In order to select the significant variables, 25 variables were derived using Lasso (least absolute shrinkage and selection operator), one of the machine learning techniques. We used Deep Learning 's LSTM technique, one of the field of machine learning for the prediction.