BayesSeg Unsupervised Segmentation for Electricity Consumption Time Series
August 4, 2026
BayesSeg integrates a dual steady-state criterion with automatic parameter optimization to partition electricity time series into steady-state and transition-state segments. It utilizes a composite metric combining event-level F1 scores and Normalized Mutual Information for multidimensional evaluation.
HOW THIS AFFECTS YOU
●
researcherYou can use this framework to automate unsupervised partitioning in Non-Intrusive Load Monitoring tasks.