Prediction of fluctuation loads based on GARCH family-CatBoost-CNNLSTM

Huibin Zeng, Bilin Shao, Hongbin Dai, Yichuan Yan and Ning Tian

Energy, 2023, vol. 263, issue PE

Abstract: An accurate prediction on natural gas load is always a guarantee of a safe and reliable operation of the natural gas pipeline network system, however, natural gas daily load variations are instability and fluctuation. Therefore, accurate prediction fluctuation loads can be challenging. Concerning that many influencing factors of natural gas daily load are non-linear and time-varying, a combined prediction model combining GARCH family models, CatBoost algorithm, CNN and LSTM is proposed in this paper. The model combines two important techniques. The first technique is to visualize the fluctuation of the daily load of natural gas through the classical GARCH family model. and characterize the fluctuation through parameters. The second one is a new gradient boosting algorithm that can take into account factors such as the parameters of the GARCH family of models and the meteorological environment, and screen for suitable and important prediction features. It also combines CNN with LSTM to predict natural gas daily load with large fluctuations and calculates 95% confidence intervals. The experimental results show that the natural gas load prediction based on GARCH family-CatBoost-CNNLSTM is reduced by an average of 26.555%, 30.892%, and 26.283% in the three relative evaluation indicators of RMSE, MAE and MAPE, and an average increase of 0.914% in R2 compared with other single or combined control models such as LSTM, CNN and MLP. The model effectively combines the advantages of econometric methods, machine learning algorithms, deep learning algorithms and other techniques, as a result, it can be better applied to the prediction of fluctuation loads.

Keywords: Fluctuation loads prediction; GARCH family models; CNN; LSTM; CatBoost (search for similar items in EconPapers)
Date: 2023
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (8)

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0360544222030110
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:energy:v:263:y:2023:i:pe:s0360544222030110

DOI: 10.1016/j.energy.2022.126125

Access Statistics for this article

Energy is currently edited by Henrik Lund and Mark J. Kaiser

More articles in Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().