An xLSTM–XGBoost Ensemble Model for Forecasting Non-Stationary and Highly Volatile Gasoline Price

High-frequency fluctuations in the international crude oil market have led to multilevel characteristics in China’s domestic refined oil pricing mechanism. To address the poor fitting performance of single deep learning models on oil price data, which hampers accurate gasoline price prediction, this...

全面介紹

Saved in:
書目詳細資料
Main Authors: Fujiang Yuan, Xia Huang, Hong Jiang, Yang Jiang, Zihao Zuo, Lusheng Wang, Yuxin Wang, Shaojie Gu, Yanhong Peng
格式: Article
語言:英语
出版: MDPI AG 2025-06-01
叢編:Computers
主題:
在線閱讀:https://www.mdpi.com/2073-431X/14/7/256
標簽: 添加標簽
沒有標簽, 成為第一個標記此記錄!
_version_ 1839616216309694464
author Fujiang Yuan
Xia Huang
Hong Jiang
Yang Jiang
Zihao Zuo
Lusheng Wang
Yuxin Wang
Shaojie Gu
Yanhong Peng
author_facet Fujiang Yuan
Xia Huang
Hong Jiang
Yang Jiang
Zihao Zuo
Lusheng Wang
Yuxin Wang
Shaojie Gu
Yanhong Peng
author_sort Fujiang Yuan
collection DOAJ
description High-frequency fluctuations in the international crude oil market have led to multilevel characteristics in China’s domestic refined oil pricing mechanism. To address the poor fitting performance of single deep learning models on oil price data, which hampers accurate gasoline price prediction, this paper proposes a gasoline price prediction method based on a combined xLSTM–XGBoost model. Using gasoline price data from June 2000 to November 2024 in Sichuan Province as a sample, the data are decomposed via STL decomposition to extract trend, residual, and seasonal components. The xLSTM model is then employed to predict the trend and seasonal components, while XGBoost predicts the residual component. Finally, the predictions from both models are combined to produce the final forecast. The experimental results demonstrate that the proposed xLSTM–XGBoost model reduces the MAE by 14.8% compared to the second-best sLSTM–XGBoost model and by 83% compared to the traditional LSTM model, significantly enhancing prediction accuracy.
format Article
id doaj-art-83aa90cdf79c4197b3dcf35e3fbdcc88
institution Matheson Library
issn 2073-431X
language English
publishDate 2025-06-01
publisher MDPI AG
record_format Article
series Computers
spelling doaj-art-83aa90cdf79c4197b3dcf35e3fbdcc882025-07-25T13:18:59ZengMDPI AGComputers2073-431X2025-06-0114725610.3390/computers14070256An xLSTM–XGBoost Ensemble Model for Forecasting Non-Stationary and Highly Volatile Gasoline PriceFujiang Yuan0Xia Huang1Hong Jiang2Yang Jiang3Zihao Zuo4Lusheng Wang5Yuxin Wang6Shaojie Gu7Yanhong Peng8School of Computer Science and Technology, Taiyuan Normal University, Jinzhong 030619, ChinaCollege of Mechanical Engineering, Chongqing University of Technology, Chongqing 400054, ChinaCollege of Mechanical Engineering, Chongqing University of Technology, Chongqing 400054, ChinaCollege of Mechanical Engineering, Chongqing University of Technology, Chongqing 400054, ChinaCollege of Mechanical Engineering, Chongqing University of Technology, Chongqing 400054, ChinaCollege of Mechanical Engineering, Chongqing University of Technology, Chongqing 400054, ChinaSchool of Energy and Power, Jiangsu University of Science and Technology, Zhenjiang 212100, ChinaMagnesium Research Center, Kumamoto University, Kumamoto 860-8555, JapanCollege of Mechanical Engineering, Chongqing University of Technology, Chongqing 400054, ChinaHigh-frequency fluctuations in the international crude oil market have led to multilevel characteristics in China’s domestic refined oil pricing mechanism. To address the poor fitting performance of single deep learning models on oil price data, which hampers accurate gasoline price prediction, this paper proposes a gasoline price prediction method based on a combined xLSTM–XGBoost model. Using gasoline price data from June 2000 to November 2024 in Sichuan Province as a sample, the data are decomposed via STL decomposition to extract trend, residual, and seasonal components. The xLSTM model is then employed to predict the trend and seasonal components, while XGBoost predicts the residual component. Finally, the predictions from both models are combined to produce the final forecast. The experimental results demonstrate that the proposed xLSTM–XGBoost model reduces the MAE by 14.8% compared to the second-best sLSTM–XGBoost model and by 83% compared to the traditional LSTM model, significantly enhancing prediction accuracy.https://www.mdpi.com/2073-431X/14/7/256xLSTMXGBoostLSTMgasoline price
spellingShingle Fujiang Yuan
Xia Huang
Hong Jiang
Yang Jiang
Zihao Zuo
Lusheng Wang
Yuxin Wang
Shaojie Gu
Yanhong Peng
An xLSTM–XGBoost Ensemble Model for Forecasting Non-Stationary and Highly Volatile Gasoline Price
Computers
xLSTM
XGBoost
LSTM
gasoline price
title An xLSTM–XGBoost Ensemble Model for Forecasting Non-Stationary and Highly Volatile Gasoline Price
title_full An xLSTM–XGBoost Ensemble Model for Forecasting Non-Stationary and Highly Volatile Gasoline Price
title_fullStr An xLSTM–XGBoost Ensemble Model for Forecasting Non-Stationary and Highly Volatile Gasoline Price
title_full_unstemmed An xLSTM–XGBoost Ensemble Model for Forecasting Non-Stationary and Highly Volatile Gasoline Price
title_short An xLSTM–XGBoost Ensemble Model for Forecasting Non-Stationary and Highly Volatile Gasoline Price
title_sort xlstm xgboost ensemble model for forecasting non stationary and highly volatile gasoline price
topic xLSTM
XGBoost
LSTM
gasoline price
url https://www.mdpi.com/2073-431X/14/7/256
work_keys_str_mv AT fujiangyuan anxlstmxgboostensemblemodelforforecastingnonstationaryandhighlyvolatilegasolineprice
AT xiahuang anxlstmxgboostensemblemodelforforecastingnonstationaryandhighlyvolatilegasolineprice
AT hongjiang anxlstmxgboostensemblemodelforforecastingnonstationaryandhighlyvolatilegasolineprice
AT yangjiang anxlstmxgboostensemblemodelforforecastingnonstationaryandhighlyvolatilegasolineprice
AT zihaozuo anxlstmxgboostensemblemodelforforecastingnonstationaryandhighlyvolatilegasolineprice
AT lushengwang anxlstmxgboostensemblemodelforforecastingnonstationaryandhighlyvolatilegasolineprice
AT yuxinwang anxlstmxgboostensemblemodelforforecastingnonstationaryandhighlyvolatilegasolineprice
AT shaojiegu anxlstmxgboostensemblemodelforforecastingnonstationaryandhighlyvolatilegasolineprice
AT yanhongpeng anxlstmxgboostensemblemodelforforecastingnonstationaryandhighlyvolatilegasolineprice
AT fujiangyuan xlstmxgboostensemblemodelforforecastingnonstationaryandhighlyvolatilegasolineprice
AT xiahuang xlstmxgboostensemblemodelforforecastingnonstationaryandhighlyvolatilegasolineprice
AT hongjiang xlstmxgboostensemblemodelforforecastingnonstationaryandhighlyvolatilegasolineprice
AT yangjiang xlstmxgboostensemblemodelforforecastingnonstationaryandhighlyvolatilegasolineprice
AT zihaozuo xlstmxgboostensemblemodelforforecastingnonstationaryandhighlyvolatilegasolineprice
AT lushengwang xlstmxgboostensemblemodelforforecastingnonstationaryandhighlyvolatilegasolineprice
AT yuxinwang xlstmxgboostensemblemodelforforecastingnonstationaryandhighlyvolatilegasolineprice
AT shaojiegu xlstmxgboostensemblemodelforforecastingnonstationaryandhighlyvolatilegasolineprice
AT yanhongpeng xlstmxgboostensemblemodelforforecastingnonstationaryandhighlyvolatilegasolineprice