ANALYSIS OF THE CLASSICAL DECOMPOSITION AND EXPONENTIAL SMOOTHING MODELS IN A HYBRID ARBOR IN THE FORECAST OF MONTHLY RESIDENTIAL ELECTRICITY CONSUMPTION

- 326238
Extended Abstracts (EA)
Favorite this paper
How to cite this paper?
Abstract

Electricity consumption forecasting plays a fundamental role in the activities of the electricity sector. This work aims to evaluate the performance of time series forecasting models applied to the prediction of monthly residential electricity consumption per consumer in Brazil, over a 12-month horizon. Two time series forecasting models were used - classical decomposition and exponential smoothing. Both models were applied individually and then in a combined (hybrid) approach, using the methods of arithmetic mean, minimum variance, weighted geometric mean, weighted harmonic mean, and Inverse Mean Square. Among the individual models, the exponential smoothing model was better able to capture the seasonality of the time series, while the classical decomposition model, despite presenting the highest correlation between the predicted and observed series, showed inferior performance. In a general analysis, the combination methods showed little significant gain over the individual models, with the Inverse Mean Square method obtaining the best performance.

Share your ideas or questions with the authors!

Did you know that the greatest stimulus in scientific and cultural development is curiosity? Leave your questions or suggestions to the author!

Sign in to interact

Have a question or suggestion? Share your feedback with the authors!

Institutions
  • 1 Centro Federal de Educação Tecnológica Celso Suckow da Fonseca – CEFET-RJ
Track
  • 7. EN&OG – OR in Energy, Oil and Gas
Keywords
Electricity consumption forecasting
Time series analysis
Hybrid methods