NEWS
tscv 1.0.2 (2026-10-21)
- Updated unit tests for compatibility with the upcoming
fabletools 1.0.0 release by using is_mable() instead of relying on the internal mdl_df class name.
tscv 1.0.1 (2026-09-02)
- Added
test_seasonality() to test for seasonality at a specified frequency using the autocorrelation-based procedure from the M4 Forecasting Competition.
- Added
forecast_naive2() to generate Naive2 benchmark forecasts using conditional seasonal adjustment.
- Added
NAIVE2() and supporting S3 methods to integrate the Naive2 benchmark with the fabletools model interface, including forecast distributions, fitted values and residuals.
- Expanded the package documentation with references covering time series cross-validation, forecast accuracy measures, and benchmark forecasting methods.
tscv 1.0.0 (2026-05-13)