gemmr.sample_size.interpolation.calc_n_required_all_metrics

gemmr.sample_size.interpolation.calc_n_required_all_metrics(ds, target_power=0.9, target_error=0.1, search_dim='n', prefix='', average_rep=True)

Calculate n_required for 5 commonly used metrics, as well as maximum across metrics.

Required sample sizes are calculated with calc_n_required().

Used metrics are power, metrics.mk_betweenAssocRelError(), metrics.mk_weightError(), metrics.mk_scoreError(), and metrics.mk_loadingError()

NOTE: removed “.sel(mode=0)” and “.dropna(‘iter’, ‘all’)”

Parameters:

  • ds (xr.Dataset) – dataset on which metrics will be computed.

  • target_power (float between 0 and 1) – minimum acceptable power

  • target_error (float between 0 and 1) – maximum accepted error

  • average_rep (bool) – if True calculated values of metrics are averaged across dimension 'rep' before calculating required sample size (doesn’t apply to metric power)

Returns:

all_n_requireds – entries are xr.DataArray s giving required sample sizes

Return type:

dict