flatness: flatness: a package to assess the flatness of (rank)...

In flatness: Indices and Tests for Assessing the Flatness of (Rank) Histograms

Description

The flatness package offers tools (scores, tests, ...) to compute histograms and assess whether they are flat.

Details

The S3 generic function rkhist allows to compute one or several rank histograms from ensemble forecasts and corresponding scalar observations. (In Meteorology an ensemble forecast is a set of forecasts for the same variable, aimed at assessing the forecasting uncertainty). It creates an object with class rkhist that can then be plotted and printed.

Flatness of (rank) histograms may then be tested with function JP_test that implements the Jolliff-Primo flatness tests. This test requires a set of deviance vectors, some of which can be provided with functions named deviate_XXX. The user can easily implement its own deviate-returning function (please see details in get_deviates on how to do this). Functions is_JP_ready and make_JP_ready are provided to ensure that a set of deviate vectors meet the requirements to be used in the Jolliffe-Primo tests. The result of the test is stored in an object with class JPtest that can be printed or drawn with the function lattice::levelplot.

Flatness indices can be computed with the S3 generic function flatness_indices.

See the vignette for further details and an illustration with the datasets ensembles and ppensembles provided with this package.

flatness documentation built on June 29, 2021, 9:08 a.m.