Evaluating multiclass models

All of the same principles that we used to evaluate binary classification models apply to multiclass model evaluation. Computing a confusion matrix is just as important, though a fair bit more difficult to interpret. We also still need to examine somewhat competing measures, such as precision and sensitivity. This, too, is messier than doing so with binary classification.

Once again, we will work with the NLS degree completion data. We will alter the target in this case, from bachelor’s degree completion or not to high school completion, bachelor’s degree completion, and post-graduate degree completion:

1. We will start by loading the necessary libraries. These are the same libraries we used in the previous two sections:

   import pandas as pd
   import numpy as np
   from feature_engine.encoding import OneHotEncoder
   from sklearn.model_selection import train_test_split
   from sklearn.preprocessing import StandardScaler
   from sklearn.neighbors...