Model 1
收藏IEEE2020-10-14 更新2026-04-17 收录
下载链接：
https://ieee-dataport.org/analysis/model-1-0
下载链接
链接失效反馈官方服务：
资源简介：
Long-term (up to 11 months ahead) forecasting using regression Ensemble Trees. For monthly (kWh) consumption forecasting, we trained twelve regression models (one per month), using (daily kWh) consumption extracted features. From weekend/weekday and full series (170 features total):Statistical: median, variance, quantiles, ...Time-series: autocorrelations, trends, seasonalities, ...Data separation:General regression learning workflow: 60% (train), 20% (validation), 20% (test).Model types:Gradient Boosted Trees (using python XGBoost)Extremes:Worst case: January forecasts (minimal samples available: only 240 smart meters).Best case: December forecasts (maximal: 3238 smart meters).Most important features:Winter & summer:Trend statistics, using the LOESS seasonal decomposition (see: https://www.statsmodels.org/stable/examples/notebooks/generated/stl_decomposition.html). Specifically 90-95% quantiles.Summer also:Weekend/weekday separated statistics.
创建时间：
2020-10-14
