Applying an Artificial Neural Network

Now that you know the components of an ANN, as well as the different steps that it follows to train a model and make predictions, let's train a simple network using the scikit-learn library.

In this topic, scikit-learn's neural network module will be used to train a network using the datasets used in the previous chapter's exercises and activities (that is, the Fertility Dataset and the Processed Census Income Dataset). It is important to mention that scikit-learn is not the most appropriate library for neural networks, as it does not currently support many types of neural networks, and its performance over deeper networks is not as good as other neural network specialized libraries, such as TensorFlow and PyTorch.

The neural network module in scikit-learn currently supports an MLP for classification, an MLP for regression, and a Restricted Boltzmann Machine architecture. Considering that the case study consists of a classification...