Getting Started with scanr
1 Overview
The scanr package detects multiple change points in univariate time series using window-based SCAN statistics. The method compares local windows along a series, calibrates evidence for changes using resampling, and combines information across several window sizes.
This book demonstrates a typical workflow:
- simulate a univariate time series with known change points;
- run
scan_cpd()with multiple window sizes; - compare estimated change points with the known truth;
- inspect the result using diagnostic visualizations;
- repeat the workflow for distributional changes;
- use the SWAL statistic for single change-point localization.