Abstract:Long-term forecasting is of paramount importance in numerous scenarios, including predicting future energy, water, and food consumption. For instance, extreme weather events and natural disasters can profoundly impact infrastructure operations and pose severe safety concerns. Traditional CNN-based models often struggle to capture long-distance dependencies effectively. In contrast, Transformers-based models have shown significant promise in long-term forecasting. This paper investigates the long-term forecasting problem and identifies a common limitation in existing Transformer-based models: they tend to reduce computational complexity at the expense of time information aggregation capability. Moreover, the order of time series plays a crucial role in accurate predictions, but current Transformer-based models lack sensitivity to time series order, rendering them unreasonable. To address these issues, we propose a novel Deformable-Local (DL) aggregation mechanism. This mechanism enhances the model's ability to aggregate time information and allows the model to adaptively adjust the size of the time aggregation window. Consequently, the model can discern more complex time patterns, leading to more accurate predictions. Additionally, our model incorporates a Frequency Selection module to reinforce effective features and reduce noise. Furthermore, we introduce Position Weights to mitigate the order-insensitivity problem present in existing methods. In extensive evaluations of long-term forecasting tasks, we conducted benchmark tests on six datasets covering various practical applications, including energy, traffic, economics, weather, and disease. Our method achieved state-of-the-art (SOTA) results, demonstrating significant improvements. For instance, on the ETT dataset, our model achieved an average MSE improvement of approximately 19% and an average MAE improvement of around 27%. Remarkably, for predicted lengths of 96 and 192, we achieved outstanding MSE and MAE improvements of 32.1% and 30.9%, respectively.

Improving Long-Term Electricity Time Series Forecasting in Smart Grid with a Three-Stage Channel-Temporal Approach

NTDformer: A Multi-Scale Forecasting Model for Non-Stationary Multilevel Time Series

Mat-Transformer-Based State Prediction Method for Information Equipment

Short-Term Electricity Load Forecasting Using the Temporal Fusion Transformer: Effect of Grid Hierarchies and Data Sources

TS2ARCformer: A Multi-Dimensional Time Series Forecasting Framework for Short-Term Load Prediction

MAT-transformer-based state forecasting method for information devices

Short-Term Load Forecasting Based on the Transformer Model

SCAT: A Time Series Forecasting with Spectral Central Alternating Transformers

Short-Term Load Forecasting for Regional Smart Energy Systems Based on Two-Stage Feature Extraction and Hybrid Inverted Transformer

Self-Attention Mechanism to Enhance the Generalizability of Data-Driven Time-Series Prediction: A Case Study of Intra-Hour Power Forecasting of Urban Distributed Photovoltaic Systems

Temporal Convolution and Multi-Attention Jointly Enhanced Electricity Load Forecasting.

Electric Load Forecasting with Fast Fourier Transform-Optimized Transformers

Physically-guided Temporal Diffusion Transformer for Long-Term Time Series Forecasting

An Ensemble Framework for Short-Term Load Forecasting Based on TimesNet and TCN

Huffman Tree Based Multi-resolution Temporal Convolution Network for Electricity Time Series Prediction

Enhancing Short-Term Power Load Forecasting With a TimesNet-Crossformer-LSTM Approach

MTS-LSTDM: Multi-Time-Scale Long Short-Term Double Memory for power load forecasting

A Comparative Study of Machine Learning-Based Time Series Prediction of Electricity Consumption in Virtual Power Plants

PWDformer: Deformable Transformer for Long-Term Series Forecasting

A Transformer Based Method with Wide Attention Range for Enhanced Short-term Load Forecasting

Short-Term Electrical Load Forecasting Based on Time Augmented Transformer