Home // GEOProcessing 2020, The Twelfth International Conference on Advanced Geographic Information Systems, Applications, and Services // View article

Electric Energy Consumption Forecast based on Spatial Information

Authors:
Carolina Cipriano
Mayara Silva
Weldson Corrêa
João Almeida
Márcia Silva
João Diniz

Keywords: Geospatial Information; Energy Forecasting; STL; ARIMA; BAQR

Abstract:
The task of predicting the consumer's electricity consumption is currently a trend in power energy companies. This prediction becomes difficult or impractical for consumers with no history or a short history of consumption. Thus, this work deals with an alternative to the prediction of energy consumption for these consumers. The proposed method is based on the consumption of the k closest neighbors and the consumption forecast made by one of three available regression models. The regressors used, namely Autoregressive Integrated Moving Average (ARIMA), Boosting Additive Quantile Regression (BAQR) and the named Seasonal and Trend decomposition using Loess (STL), were chosen for providing the best performance. The results obtained were promising, achieved a mean of the 30.4 % in the symmetric mean absolute percentage error (sMAPE) metric in a dataset with 86,874 customers.

Pages: 29 to 35

Copyright: Copyright (c) IARIA, 2020

Publication date: March 22, 2020

Published in: conference

ISSN: 2308-393X

ISBN: 978-1-61208-762-7

Location: Valencia, Spain

Dates: from November 21, 2020 to November 25, 2020