Real-data validation of a window-based nonlinear time-series analysis framework using electricity load data

This archive provides the real-world electricity load datasets and window-level analysis results used to validate a window-based nonlinear time-series analysis framework. The original electricity load data are from the UCI Machine Learning Repository (Electricity Load Diagrams 2011–2014). Processed multiscale datasets and residual-based window scores are provided together with a script to reproduce the results. This archive is intended to support methodological reproducibility and robustness assessment. No domain-specific modeling or forecasting of electricity systems is performed.