Checking model fit
In the video you analyzed the example of an improvement in the model fit by adding additional variable on the wells
data. Continuing with this data set you will see how further increase in model complexity effects deviance and model fit.
The dataset wells
have been preloaded in the workspace.
This exercise is part of the course
Generalized Linear Models in Python
Exercise instructions
- Fit a logistic regression model with
switch
as the response anddistance100
andarsenic
as explanatory variables. - Compute the difference in deviance of the intercept only model and the model including all the variables.
- Print the computed difference.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Import statsmodels
import ____.____ as sm
from ____.____.____ import glm
# Define model formula
formula = '____ ~ ____'
# Fit GLM
model_dist_ars = glm(____, ____ = ____, ____ = sm.____.____).____
# Compare deviance of null and residual model
diff_deviance = ____.____ - ____.____
# Print the computed difference in deviance
____(____)