APPLYING N-BEATS MODEL FOR SHORT-TERM LOAD FORECASTING IN VIETNAM

  • Nguyễn Anh Tuấn
  • Lê Anh Ngọc

Abstract

In the energy industry in general, load forecasting is extremely important because electricity is directly affecting daily life and national economic sectors. Currently, there are many load forecasting methods, and each method has different its own advantages and disadvantages. Unlike current deep learning models, N-BEATS model has the advantages of fast training and inference time. The model also achieved high performance in famous time-series prediction competitions in different fields. In addition, the model can interpret the trend and seasonality of the data. This paper presents a shortterm (specifically 48 hours) electricity load forecasting method in Viet Nam based on the N-BEATS deep learning model. The paper analyzes the results and effects of trend and seasonality in load time series.

điểm /   đánh giá
Published
2022-10-18
Section
Bài viết