D2Vformer: A Flexible Time-Series Prediction Model Based on Time-Position Embedding
- 4 December 2025
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks and Learning Systems
- Vol. 37 (5) , 2223-2234
Abstract
Existing time-series forecasting methods often struggle to adapt to dynamic scenarios and lack flexibility in prediction. They typically require retraining the model when the prediction length or position changes. Moreover, these methods still face challenges in effectively capturing and utilizing time-position embeddings (PEs). To address these limitations, this article proposes a novel model called D2Vformer. Unlike conventional prediction methods that rely on fixed-length predictors, D2Vformer can directly handle scenarios with arbitrary prediction lengths. In addition, it significantly reduces training resource consumption and proves highly effective in real-world dynamic environments. In D2Vformer, the Date2Vec (D2V) module is devised to leverage timestamp information and feature sequences to generate time PEs. Subsequently, D2Vformer introduces an innovative fusion module that leverages an attention mechanism to capture the mapping between input and target time PEs, thereby enabling flexible prediction. Extensive experiments on six datasets demonstrate that D2V outperforms other time-PE methods, while D2Vformer surpasses state-of-the-art approaches in both fixed-length and arbitrary-length prediction tasks. The code for D2Vformer is available at: https://github.com/TeamofHaoWang/D2VformerKeywords
All Related Versions
Funding Information
- National Natural Science Foundation of China (62206178)
- National Natural Science Foundation of China (72301180)
- Natural Science Foundation of Guangdong Province (2025A1515012015)
- Stable Support Plan for Higher Education Institutions in Shenzhen (20231121221536001)
This publication has 19 references indexed in Scilit:
- Probabilistic Approximation of Stochastic Time Series Using Bayesian Recurrent Neural NetworkIEEE Transactions on Neural Networks and Learning Systems, 2025
- Deep learning-based time series forecastingArtificial Intelligence Review, 2024
- Denoising Multiscale Spectral Graph Wavelet Neural Networks for Gas Utilization Ratio Prediction in Blast FurnaceIEEE Transactions on Neural Networks and Learning Systems, 2024
- A Continuous Volatility Forecasting Model Based on Neural Differential Equations and Scale-SimilarityIEEE Transactions on Neural Networks and Learning Systems, 2024
- A Novel Sequence-to-Sequence-Based Deep Learning Model for Multistep Load ForecastingIEEE Transactions on Neural Networks and Learning Systems, 2024
- Is Single Enough? A Joint Spatiotemporal Feature Learning Framework for Multivariate Time Series PredictionIEEE Transactions on Neural Networks and Learning Systems, 2022
- Building Trend Fuzzy Granulation-Based LSTM Recurrent Neural Network for Long-Term Time-Series ForecastingIEEE Transactions on Fuzzy Systems, 2021
- Deep learning with long short-term memory networks for financial market predictionsEuropean Journal of Operational Research, 2018
- A hybrid forecasting approach applied to wind speed time seriesRenewable Energy, 2013
- Evaluation of hybrid forecasting approaches for wind speed and power generation time seriesRenewable and Sustainable Energy Reviews, 2012