Make a dataset class-imbalanced

An illustration of the make_imbalance() function to create an imbalanced dataset from a balanced dataset. We show the ability of make_imbalance() of dealing with Pandas DataFrame.

# Adapted from imbalanced-learn
# Authors: Dayvid Oliveira
#          Christos Aridas
#          Guillaume Lemaitre
#          Zhining Liu <zhining.liu@outlook.com>
# License: MIT

print(__doc__)

import matplotlib.pyplot as plt
import seaborn as sns
sns.set_context("poster")

Generate the dataset

First, we will generate a dataset and convert it to a DataFrame with arbitrary column names. We will plot the original dataset.

import pandas as pd
from sklearn.datasets import make_moons

X, y = make_moons(n_samples=200, shuffle=True, noise=0.25, random_state=10)
X = pd.DataFrame(X, columns=["feature 1", "feature 2"])

fig = plt.figure(figsize=(6, 5))
ax = sns.scatterplot(
    data=X,
    x="feature 1",
    y="feature 2",
    hue=y,
    style=y,
)

Make a dataset imbalanced

Now, we will show the helpers make_imbalance() that is useful to random select a subset of samples. It will impact the class distribution as specified by the parameters.

from collections import Counter


def ratio_func(y, multiplier, minority_class):
    target_stats = Counter(y)
    return {minority_class: int(multiplier * target_stats[minority_class])}

from imbens.datasets import make_imbalance

fig, axs = plt.subplots(nrows=2, ncols=3, figsize=(15, 10))

sns.scatterplot(
    data=X,
    x="feature 1",
    y="feature 2",
    hue=y,
    style=y,
    ax=axs[0, 0],
)
axs[0, 0].set_title("Original set")

multipliers = [0.9, 0.75, 0.5, 0.25, 0.1]
for ax, multiplier in zip(axs.ravel()[1:], multipliers):
    X_resampled, y_resampled = make_imbalance(
        X,
        y,
        sampling_strategy=ratio_func,
        **{"multiplier": multiplier, "minority_class": 1},
    )

    sns.scatterplot(
        data=X_resampled,
        x="feature 1",
        y="feature 2",
        hue=y_resampled,
        style=y_resampled,
        ax=ax,
    )
    ax.set_title(f"Sampling ratio = {multiplier}")

plt.tight_layout()
plt.show()