pythonstatsmodelsfstat

How to extract the F statistic and the P value from results.f_test in statsmodels python


I run the OLS.fit() and use the fitted model to get the f value for a particular category like in ANCOVA.

res = OLS.fit()
R = [[0,1,0], [0,0,1]]
res.f_test(R)

The result is in the form of

<F test: F=array([[ 2.21065884]]), p=[[ 0.10971892]], df_denom=5918, df_num=2>

How would I be able to get the F array and the p value from this format? or Is it even possible?


Solution

  • Assign the results to store it and then check the attributes.

    fres = res.f_test(R)
    dir(fres)
    fres.fvalue
    fres.pvalue
    

    or similar, since I am working with statsmodels master

    http://statsmodels.sourceforge.net/devel/generated/statsmodels.regression.linear_model.RegressionResults.f_test.html