On Alberta's 42-area load data, global LightGBM and XGBoost with weighted instance clustering outperform local models, while global Ridge regression is slightly less accurate but far more scalable.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Globalization for Scalable Short-term Load Forecasting
On Alberta's 42-area load data, global LightGBM and XGBoost with weighted instance clustering outperform local models, while global Ridge regression is slightly less accurate but far more scalable.