Multidomain Electrical Load Forecasting Benchmark Datasets
收藏资源简介:
This dataset collection supports the study “Predictability of Electrical Loads in Various Domains: An Empirical Benchmarking Approach”. It contains cleaned and feature-enriched time series for four energy consumption domains — Residential, Commercial, Industrial, and Aggregated Loads — excluding proprietary Fraunhofer data. All datasets cover a two-year period with a 15-minute resolution and include relevant weather and calendar-based features (e.g., temperature, humidity, wind speed, holidays). Residential data: 20 selected single-family load profiles from the WPuQ dataset (Germany). Commercial and industrial data: 20 and 10 profiles respectively, selected via clustering from the EWELD dataset (China). Aggregated data: 20 time series from national (Europe), regional (USA), and local (Australia) sources.



