Preparing the data

We should do a few more data checks. Most importantly, let’s check the balance between classes in the training data:

1. Start with the following code:

   clf_df['label'].value_counts()

   …

   0    66

   1    11

   Name: label, dtype: int64

The data is imbalanced, but not too badly.

2. Let’s get this in percentage form, just to make this a little easier to understand:

   clf_df['label'].value_counts(normalize=True)

   …

   0    0.857143

   1    0.142857

   Name: label, dtype: float64

It looks like we have about an 86/14 balance between the classes. Not awful. Keep this in mind, because the model should be able to predict with about 86% accuracy just based on the imbalance alone. It won’t be an impressive model at all if it only hits 86%.

3. Next, we need to cut up our data for our model. We will use the features as our X data, and the answers...