statsmodels.tsa.stattools.ClarkWestResult#

class statsmodels.tsa.stattools.ClarkWestResult(statistic, pvalue, lags, kernel, mspe_restricted, mspe_unrestricted, adjustment)[source]#

Result of clark_west_test.

Added in version 0.15.1.

Parameters:
statisticfloat

The Clark-West test statistic, approximately standard normal under the null of equal predictive accuracy.

pvaluefloat

The one-sided p-value against the alternative that the unrestricted model is more accurate, from the standard normal distribution.

lagsint

The number of lags used in the HAC estimator of the long-run variance of the adjusted loss differential.

kernelstr

The kernel of the HAC estimator, “bartlett” or “uniform”.

mspe_restrictedfloat

The mean squared prediction error of forecast_restricted.

mspe_unrestrictedfloat

The mean squared prediction error of forecast_unrestricted.

adjustmentfloat

The mean squared difference between the two forecasts. The mean of the adjusted loss differential is mspe_restricted - (mspe_unrestricted - adjustment).

Notes

Unpacks as statistic, pvalue = result. The other fields are only available through attribute access.

Methods

Properties