The default behavior for adding and solving with noisemodels has changed from Pastas 1.5. Find more information here

pastas.stats.metrics#

The following methods may be used to describe the fit between the model simulation and the observations.

Examples

These methods may be used as follows:

>>> ps.stats.rmse(sim, obs)

or directly from a Pastas model:

>>> ml.stats.rmse()

Functions

aic

Compute the Akaike Information Criterium (AIC).

aicc

Compute the Akaike Information Criterium with second order bias correction for the number of observations (AICc)

bic

Compute the Bayesian Information Criterium (BIC).

evp

Compute the (weighted) Explained Variance Percentage (EVP).

kge

Compute the (weighted) Kling-Gupta Efficiency (KGE).

kge_2012

Compute the (weighted) Kling-Gupta Efficiency (KGE).

mae

Compute the (weighted) Mean Absolute Error (MAE).

nse

Compute the (weighted) Nash-Sutcliffe Efficiency (NSE).

pearsonr

Compute the (weighted) Pearson correlation (r).

rmse

Compute the (weighted) Root Mean Squared Error (RMSE).

rsq

Compute R-squared, possibly adjusted for the number of free parameters.

sse

Compute the Sum of the Squared Errors (SSE).