Now that we have some neural network training functionality that appears to be working, let's try to utilize this functionality in a more realistic modeling scenario. In particular, let's bring back our favorite classification dataset, the iris flower dataset (utilized in Chapter 5, Classification).
If you remember, when trying to classify iris flowers using this dataset, we are trying to classify them into one of three species (setosa, virginica, or versicolor). As our neural net is expecting matrices of float values, we need to encode the three species into numerical columns. One way to do this is to create a column in our dataset for each species. We will then set that column's values to either 1.0 or 0.0 depending on whether the corresponding row's measurements correspond to that species (1.0) or to another species (0...
