tests

Statistical tests to check model results.

Contains diagnostic tests to check if model residuals/noise do not contain significant autocorrelation and are normally distributed, using hypothesis testing.

Examples

Run diagnostic tests on model after optimization:

ml.stats.diagnostics()

Functions

durbin_watson(→ float)	Durbin-Watson test for autocorrelation.
ljung_box(→ tuple[float, float])	Ljung-box test for autocorrelation.
runs_test(→ tuple[float, float])	Runs test for autocorrelation.
stoffer_toloi(→ tuple[float, float])	Adapted Ljung-Box test to deal with missing data [Rcb3cfb20bb89-stoffer_1992].
diagnostics(, float_fmt)	Methods to compute various diagnostics checks for a time series.