probnumeval.timeseries

Evaluate time-series algorithms.

This package offers methods that evaluate the efficiency, accuracy, and calibration of time-series algorithms such as the filters/smoothers and differential equation solvers in ProbNum.

Functions

anees(approximate_solution, …)	Compute the average normalised estimation error squared.
non_credibility_index(approximate_solution, …)	Compute the non-credibility index (NCI).
inclination_index(approximate_solution, …)	Compute the inclination index (II).
rmse(approximate_solution, …)	Compute the root mean-square error.
relative_rmse(approximate_solution, …)	Compute the root mean-square error.
mae(approximate_solution, …)	Compute the root mean-square error.
relative_mae(approximate_solution, …)	Compute the root mean-square error.
max_error(approximate_solution, …)	Compute the root mean-square error.
relative_max_error(approximate_solution, …)	Compute the root mean-square error.
mean_error(approximate_solution, …)	Compute the mean error.
relative_mean_error(approximate_solution, …)	Compute the relative mean error.