Skip to contents

tscv 1.0.2

  • 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

CRAN release: 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

CRAN release: 2026-05-13

  • Initial CRAN submission.