Quickstart¶
Installation¶
Install with pip
:
pip install targeted
Note
Presently binary wheels are available for Linux systems and Mac OS X running Python>=3.6. Windows installations must be built from source.
Risk regression¶
As an illustration data is loaded from the package
In [1]: import targeted as tg
In [2]: d = tg.getdata() # returns a Pandas DataFrame
In [3]: print(d.head())
y a x z
0 0 0 -0.626454 1.134965
1 0 0 0.183643 1.111932
2 0 0 -0.835629 -0.870778
3 1 0 1.595281 0.210732
4 1 1 0.329508 0.069396
Here y
is the binary response variable, a
binary the exposure,
and x
, z
covariates.
Next we estimate the risk difference of the exposed vs the non-exposed
In [4]: import numpy as np
In [5]: from patsy import dmatrices
In [6]: y, X = dmatrices('y ~ x+z', d)
In [7]: a = d['a']
In [8]: ones = np.ones((y.size,1))
In [9]: tg.riskreg(y=y, a=a, x1=ones, x2=X)
Out[9]: Riskreg. Estimate: [1.02819001]
Or using the formula syntax:
In [10]: from targeted.formula import riskreg
In [11]: riskreg(d, 'y~a')
Out[11]: Riskreg. Estimate: [1.49101228]