Continuous Statistics Functions

Return to Homepage

Continuous Statistics Functions From Error Based Agreement Maps.

gval.statistics.continuous_stat_funcs.coefficient_of_determination(error: DataArray | Dataset, benchmark_map: DataArray | Dataset) → Number

Compute coefficient of determination (R2).

Either (error and benchmark_map) or (candidate_map and benchmark_map) must be provided.

Parameters:

  • error (Union[xr.DataArray, xr.Dataset]) – Candidate minus benchmark error.

  • benchmark_map (Union[xr.DataArray, xr.Dataset]) – Benchmark map.

Returns:

R2 – Coefficient of determination.

Return type:

Number

References

gval.statistics.continuous_stat_funcs.mean_absolute_error(error: DataArray | Dataset) → Number

Compute mean absolute error (MAE).

Either error or (candidate_map and benchmark_map) must be provided.

Parameters:

error (Union[xr.DataArray, xr.Dataset]) – Candidate minus benchmark error.

Returns:

MAE – Mean absolute error.

Return type:

Number

References

gval.statistics.continuous_stat_funcs.mean_absolute_percentage_error(error: DataArray | Dataset, benchmark_map: DataArray | Dataset) → Number

Compute mean absolute percentage error (MAPE).

Either (error and benchmark_map) or (candidate_map and benchmark_map) must be provided.

Parameters:

  • error (Union[xr.DataArray, xr.Dataset]) – Candidate minus benchmark error.

  • benchmark_map (Union[xr.DataArray, xr.Dataset]) – Benchmark map.

Returns:

MAPE – Mean absolute percentage error.

Return type:

Number

References

gval.statistics.continuous_stat_funcs.mean_normalized_mean_absolute_error(error: DataArray | Dataset, benchmark_map: DataArray | Dataset) → Number

Compute mean normalized mean absolute error (NMAE).

Either (error and benchmark_map) or (candidate_map and benchmark_map) must be provided.

Parameters:

  • error (Union[xr.DataArray, xr.Dataset]) – Candidate minus benchmark error.

  • benchmark_map (Union[xr.DataArray, xr.Dataset]) – Benchmark map.

Returns:

NMAE – Normalized mean absolute error.

Return type:

Number

References

gval.statistics.continuous_stat_funcs.mean_normalized_root_mean_squared_error(error: DataArray | Dataset, benchmark_map: DataArray | Dataset) → Number

Compute mean normalized root mean squared error (NRMSE).

Either (error and benchmark_map) or (candidate_map and benchmark_map) must be provided.

Parameters:

  • error (Union[xr.DataArray, xr.Dataset]) – Candidate minus benchmark error.

  • benchmark_map (Union[xr.DataArray, xr.Dataset]) – Benchmark map.

Returns:

mNRMSE – Mean normalized root mean squared error.

Return type:

Number

References

gval.statistics.continuous_stat_funcs.mean_percentage_error(error: DataArray | Dataset, benchmark_map: DataArray | Dataset) → Number

Compute mean percentage error (MPE).

Either (error and benchmark_map) or (candidate_map and benchmark_map) must be provided.

Parameters:

  • error (Union[xr.DataArray, xr.Dataset]) – Candidate minus benchmark error.

  • benchmark_map (Union[xr.DataArray, xr.Dataset]) – Benchmark map.

Returns:

MPE – Mean percentage error.

Return type:

Number

References

gval.statistics.continuous_stat_funcs.mean_signed_error(error: DataArray | Dataset) → Number

Compute mean signed error (MSiE).

Either error or (candidate_map and benchmark_map) must be provided.

Parameters:

error (Union[xr.DataArray, xr.Dataset]) – Candidate minus benchmark error.

Returns:

MSiE – Mean signed error.

Return type:

Number

References

gval.statistics.continuous_stat_funcs.mean_squared_error(error: DataArray | Dataset) → Number

Compute mean squared error (MSE).

Either error or (candidate_map and benchmark_map) must be provided.

Parameters:

error (Union[xr.DataArray, xr.Dataset]) – Candidate minus benchmark error.

Returns:

MSE – Mean squared error.

Return type:

Number

References

gval.statistics.continuous_stat_funcs.range_normalized_mean_absolute_error(error: DataArray | Dataset, benchmark_map: DataArray | Dataset) → Number

Compute range normalized mean absolute error (RNMAE).

Either (error and benchmark_map) or (candidate_map and benchmark_map) must be provided.

Parameters:

  • error (Union[xr.DataArray, xr.Dataset]) – Candidate minus benchmark error.

  • benchmark_map (Union[xr.DataArray, xr.Dataset]) – Benchmark map.

Returns:

rNMAE – Range normalized mean absolute error.

Return type:

Number

References

gval.statistics.continuous_stat_funcs.range_normalized_root_mean_squared_error(error: DataArray | Dataset, benchmark_map: DataArray | Dataset) → Number

Compute range normalized root mean squared error (RNRMSE).

Either (error and benchmark_map) or (candidate_map and benchmark_map) must be provided.

Parameters:

  • error (Union[xr.DataArray, xr.Dataset]) – Candidate minus benchmark error.

  • benchmark_map (Union[xr.DataArray, xr.Dataset]) – Benchmark map.

Returns:

rNRMSE – Range normalized root mean squared error.

Return type:

Number

References

gval.statistics.continuous_stat_funcs.root_mean_squared_error(error: DataArray | Dataset) → Number

Compute root mean squared error (RMSE).

Either error or (candidate_map and benchmark_map) must be provided.

Parameters:

error (Union[xr.DataArray, xr.Dataset]) – Candidate minus benchmark error.

Returns:

RMSE – Root mean squared error.

Return type:

Number

References

gval.statistics.continuous_stat_funcs.symmetric_mean_absolute_percentage_error(error: DataArray | Dataset, candidate_map: DataArray | Dataset, benchmark_map: DataArray | Dataset) → Number

Compute symmetric mean absolute percentage error (sMAPE).

Both candidate_map and benchmark_map must be provided. error can be provided to avoid recomputing it.

Parameters:

  • error (Union[xr.DataArray, xr.Dataset]) – Candidate minus benchmark error.

  • candidate_map (Union[xr.DataArray, xr.Dataset]) – Candidate map.

  • benchmark_map (Union[xr.DataArray, xr.Dataset]) – Benchmark map.

Returns:

sMAPE – Symmetric mean absolute percentage error.

Return type:

Number

References