AUC important but easy to understand why

[guillaume-vignal]

In some cases, the difference comes from interactions between variables and not from a single variable.

import numpy as np
import pandas as pd

df1 = pd.concat(
    [
        pd.DataFrame(
            np.zeros((10000, 2)),
            columns=["A", "B"],
        ),
        pd.DataFrame(
            np.ones((10000, 2)),
            columns=["A", "B"],
        )
    ]
)

df2 = pd.DataFrame()
df2["A"] = df1["A"].sample(frac=1).reset_index(drop=True)
df2["B"] = df1["B"].sample(frac=1).reset_index(drop=True)

An interaction plot would be helpful to understand what's going on behind the scenes.