pyfirth

The pyfirth package implements a very basic Firth-penalized logistic regression model for rare event data. There are likely more efficient and versatile methods out there.

Dependencies

1. numpy
2. scipy
3. pandas
4. statsmodels

Installation

The software package can be installed using pip by running the following command: pip install pyfirth

Example

Here is a simple example made by binarizing some continuous data to create rare event data. Note, the Firth-penalized model is slightly more conservative.

    from pyfirth.PyFirth import PyFirth 
    import statsmodels.api as sm

    dta = sm.datasets.fair.load_pandas().data

    #create a rare count dataset by binning affair time into extreme not extreme
    dta["extreme_affairs"] = (dta["affairs"] > 10.0).astype(float)
    dta = sm.add_constant(dta)
    logit_mod = sm.Logit(dta["extreme_affairs"], dta[['const','occupation', 'educ', 'occupation_husb','rate_marriage','age','yrs_married','children','religious']])
    fitted_mod=logit_mod.fit()
    print('Logistic Regression Marriage Rating (Beta, SE, P): {0:.2f}, {1:.2f}, {2:.2e}'.format(fitted_mod.params['rate_marriage'],fitted_mod.bse['rate_marriage'],fitted_mod.pvalues['rate_marriage']))
    
    firth_test=PyFirth(dta,['const','occupation', 'educ', 'occupation_husb','rate_marriage','age','yrs_married','children','religious'],'extreme_affairs',hasconst=True)
    firth_output=firth_test.fit('rate_marriage')
    print('Firth Logistic Regression Marriage Rating (Beta, SE, P): {0:.2f}, {1:.2f}, {2:.2e}'.format(firth_output['ParamTable'].loc['rate_marriage']['BETA'],firth_output['ParamTable'].loc['rate_marriage']['SE'],firth_output['PVal']))

Optimization terminated successfully.
         Current function value: 0.037929
         Iterations 12
Logistic Regression Marriage Rating (Beta, SE, P): -0.79, 0.13, 3.85e-10
Firth Logistic Regression Marriage Rating (Beta, SE, P): -0.78, 0.12, 2.17e-09