Presentation information

General Session

General Session » GS-2 Machine learning

[2G3-GS-2e] 機械学習:予測

Wed. Jun 9, 2021 1:20 PM - 3:00 PM Room G (GS room 2)

座長:竹岡 邦紘(NEC)

1:20 PM - 1:40 PM

[2G3-GS-2e-01] Studies of validity in forecasting maximum power demand by combining multiple methods

〇Hideaki Sasaki1, Shoichi Urano1 (1. Meiji University)

Keywords:Machine learning, Electricity demand forecast, Multiple regression model, Random forest, Neural network

The authors have combined a multiple regression model of statistical methods and a random forest of machine learning methods and applied to a seasonal model to forecast power demand. Therefore, this time, we predicted the power demand by a neural network, and confirmed the validity of the proposed method by comparing and verifying it with the proposed method. Furthermore, in the proposed method, effective explanatory variables were selected, learning was improved, and improvement in prediction accuracy was confirmed. In this study, using one method and one modeling period is considered as a risk, so the risk was reduced by combining methods with different prediction characteristics and multiple modeling periods.

Authentication for paper PDF access

A password is required to view paper PDFs. If you are a registered participant, please log on the site from Participant Log In.
You could view the PDF with entering the PDF viewing password bellow.