statsmodels.regression.linear_model.OLSResults.get_robustcov_results¶

OLSResults.
get_robustcov_results
(cov_type='HC1', use_t=None, **kwargs)¶ Create new results instance with robust covariance as default.
 Parameters
 cov_type
str
The type of robust sandwich estimator to use. See Notes below.
 use_tbool
If true, then the t distribution is used for inference. If false, then the normal distribution is used. If use_t is None, then an appropriate default is used, which is True if the cov_type is nonrobust, and False in all other cases.
 **kwargs
Required or optional arguments for robust covariance calculation. See Notes below.
 cov_type
 Returns
RegressionResults
This method creates a new results instance with the requested robust covariance as the default covariance of the parameters. Inferential statistics like pvalues and hypothesis tests will be based on this covariance matrix.
Notes
The following covariance types and required or optional arguments are currently available:
‘fixed scale’ uses a predefined scale
scale
: float, optionalArgument to set the scale. Default is 1.
‘HC0’, ‘HC1’, ‘HC2’, ‘HC3’: heteroscedasticity robust covariance
no keyword arguments
‘HAC’: heteroskedasticityautocorrelation robust covariance
maxlag
integer, requirednumber of lags to use
kernel
{callable, str}, optionalkernels currently available kernels are [‘bartlett’, ‘uniform’], default is Bartlett
use_correction
: bool, optionalIf true, use small sample correction
‘cluster’: clustered covariance estimator
groups
array_like[int], required :Integervalued index of clusters or groups.
use_correction
: bool, optionalIf True the sandwich covariance is calculated with a small sample correction. If False the sandwich covariance is calculated without small sample correction.
df_correction
: bool, optionalIf True (default), then the degrees of freedom for the inferential statistics and hypothesis tests, such as pvalues, f_pvalue, conf_int, and t_test and f_test, are based on the number of groups minus one instead of the total number of observations minus the number of explanatory variables. df_resid of the results instance is also adjusted. When use_t is also True, then pvalues are computed using the Student’s t distribution using the corrected values. These may differ substantially from pvalues based on the normal is the number of groups is small. If False, then df_resid of the results instance is not adjusted.
‘hacgroupsum’: Driscoll and Kraay, heteroscedasticity and autocorrelation robust covariance for panel data # TODO: more options needed here
time
array_like, requiredindex of time periods
maxlag
integer, requirednumber of lags to use
kernel
{callable, str}, optionalThe available kernels are [‘bartlett’, ‘uniform’]. The default is Bartlett.
use_correction
{False, ‘hac’, ‘cluster’}, optionalIf False the the sandwich covariance is calculated without small sample correction. If use_correction = ‘cluster’ (default), then the same small sample correction as in the case of covtype=’cluster’ is used.
df_correction
bool, optionalThe adjustment to df_resid, see cov_type ‘cluster’ above
‘hacpanel’: heteroscedasticity and autocorrelation robust standard errors in panel data. The data needs to be sorted in this case, the time series for each panel unit or cluster need to be stacked. The membership to a time series of an individual or group can be either specified by group indicators or by increasing time periods. One of
groups
ortime
is required. # TODO: we need more options heregroups
array_like[int]indicator for groups
time
array_like[int]index of time periods
maxlag
int, requirednumber of lags to use
kernel
{callable, str}, optionalAvailable kernels are [‘bartlett’, ‘uniform’], default is Bartlett
use_correction
{False, ‘hac’, ‘cluster’}, optionalIf False the sandwich covariance is calculated without small sample correction.
df_correction
bool, optionalAdjustment to df_resid, see cov_type ‘cluster’ above
Reminder:
use_correction
in “hacgroupsum” and “hacpanel” is not bool, needs to be in {False, ‘hac’, ‘cluster’}.