Mitigating Long-Term Forecasting Bias in Time-Series Neural Networks via Ensemble of Short-Term Dependencies

Wang, Jiahui; Zhou, Wenqian; Chen, Fangshu; Wang, Liming; Pan, Ruijun; Yu, Chengcheng

doi:10.3390/app15116371

Open AccessArticle

Mitigating Long-Term Forecasting Bias in Time-Series Neural Networks via Ensemble of Short-Term Dependencies

by

Jiahui Wang

,

Wenqian Zhou

,

Fangshu Chen

^*,

Liming Wang

,

Ruijun Pan

and

Chengcheng Yu

School of Computer and Information Engineering, Shanghai Polytechnic University, Shanghai 201209, China

^*

Author to whom correspondence should be addressed.

Appl. Sci. 2025, 15(11), 6371; https://doi.org/10.3390/app15116371

Submission received: 27 April 2025 / Revised: 3 June 2025 / Accepted: 4 June 2025 / Published: 5 June 2025

Download

Browse Figures

Versions Notes

Abstract

Time-series forecasting is essential for predicting future trends based on historical data, with significant applications in meteorology, transportation, and finance. However, existing models often exhibit unsatisfactory performance in long-term forecasting scenarios. To address this limitation, we propose the Time-Series Neural Networks via Ensemble of Short-Term Dependencies (TSNN-ESTD). This model leverages iTransformer as the base predictor to simultaneously train short-term and long-term forecasting models. The vanilla iTransformer’s linear decoding layer is optimized by replacing it with an LSTM layer, and an additional long-term model is introduced to enhance stability. The ensemble strategy employs short-term predictions to correct the bias in long-term forecasts. Our extensive experiments demonstrate that TSNN-ESTD reduces the MSE and MAE by 9.17% and 2.3% on five benchmark datasets.

Keywords: time-series forecasting; time-series neural networks; long-term bias; short-term dependency; ensemble learning

Share and Cite

MDPI and ACS Style

Wang, J.; Zhou, W.; Chen, F.; Wang, L.; Pan, R.; Yu, C. Mitigating Long-Term Forecasting Bias in Time-Series Neural Networks via Ensemble of Short-Term Dependencies. Appl. Sci. 2025, 15, 6371. https://doi.org/10.3390/app15116371

AMA Style

Wang J, Zhou W, Chen F, Wang L, Pan R, Yu C. Mitigating Long-Term Forecasting Bias in Time-Series Neural Networks via Ensemble of Short-Term Dependencies. Applied Sciences. 2025; 15(11):6371. https://doi.org/10.3390/app15116371

Chicago/Turabian Style

Wang, Jiahui, Wenqian Zhou, Fangshu Chen, Liming Wang, Ruijun Pan, and Chengcheng Yu. 2025. "Mitigating Long-Term Forecasting Bias in Time-Series Neural Networks via Ensemble of Short-Term Dependencies" Applied Sciences 15, no. 11: 6371. https://doi.org/10.3390/app15116371

APA Style

Wang, J., Zhou, W., Chen, F., Wang, L., Pan, R., & Yu, C. (2025). Mitigating Long-Term Forecasting Bias in Time-Series Neural Networks via Ensemble of Short-Term Dependencies. Applied Sciences, 15(11), 6371. https://doi.org/10.3390/app15116371

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Menu

Mitigating Long-Term Forecasting Bias in Time-Series Neural Networks via Ensemble of Short-Term Dependencies

Abstract

Share and Cite

Article Metrics

Article Access Statistics

Further Information

Guidelines

MDPI Initiatives

Follow MDPI