Getting ready

As logistic regression is a linear classifier, it assumes linearity in independent variables and log odds. Thus, in scenarios where independent features are linear-dependent on log odds, the model performs very well. Higher-order features can be included in the model to capture nonlinear behavior. Let's see how to build logistic regression models using major deep learning packages as discussed in the previous chapter. Internet connectivity will be required to download the dataset from the UCI repository.