Advances in Time Series Forecasting and Renewable Energy Prediction

The field of time series forecasting is witnessing significant advancements with the introduction of novel transformer-based architectures that can effectively model both temporal and inter-channel dependencies. These developments have far-reaching implications for multivariate time-series forecasting, particularly in applications such as renewable energy prediction and global air pollution forecasting. Noteworthy papers in this area include the Sentinel model, which achieves state-of-the-art performance in time series forecasting by leveraging a multi-patch attention mechanism. Another significant contribution is the probabilistic net load forecasting framework utilizing an enhanced conditional diffusion model, which demonstrates exceptional potential in uncertainty quantification for scenario forecasting. Additionally, the WeatherMesh-3 system showcases a fast and accurate operational global weather forecasting system, pushing the boundaries of machine learning-based weather prediction. The Seasonal-Periodic Decomposition Network (SPDNet) is also a notable mention, as it effectively captures intricate temporal dynamics in residential electricity demand forecasting.

Sources

Sentinel: Multi-Patch Transformer with Temporal and Channel Attention for Time Series Forecasting

Probabilistic Net Load Forecasting for High-Penetration RES Grids Utilizing Enhanced Conditional Diffusion Model

Renewable Energy Transition in South America: Predictive Analysis of Generation Capacity by 2050

Offline Meteorology-Pollution Coupling Global Air Pollution Forecasting Model with Bilinear Pooling

Leveraging Land Cover Priors for Isoprene Emission Super-Resolution

Data-driven Mesoscale Weather Forecasting Combining Swin-Unet and Diffusion Models

Long-Term Electricity Demand Prediction Using Non-negative Tensor Factorization and Genetic Algorithm-Driven Temporal Modeling

WeatherMesh-3: Fast and accurate operational global weather forecasting

SPDNet: Seasonal-Periodic Decomposition Network for Advanced Residential Demand Forecasting

Built with on top of