clustering_gmm.ipynb

Kernel: Python 3

Clustering with Gaussian mixture models

# Standard Python libraries
from __future__ import absolute_import, division, print_function, unicode_literals

import os
import time
import numpy as np
import glob
import matplotlib.pyplot as plt
import PIL
import imageio

from IPython import display

import sklearn

import seaborn as sns

sns.set(style="ticks", color_codes=True)

import pandas as pd

pd.set_option("precision", 2)  # 2 decimal places
pd.set_option("display.max_rows", 20)
pd.set_option("display.max_columns", 30)
pd.set_option("display.width", 100)  # wide windows

In this section, we show how to find clusters in an unlabeled 2d version of the Iris dataset by fitting a GMM using sklearn.

import seaborn as sns

from sklearn.datasets import load_iris

# from sklearn.cluster import KMeans
from sklearn.mixture import GaussianMixture

iris = load_iris()
X = iris.data

fig, ax = plt.subplots()
idx1 = 2
idx2 = 3
ax.scatter(X[:, idx1], X[:, idx2], c="k", marker=".")
ax.set(xlabel=iris.feature_names[idx1])
ax.set(ylabel=iris.feature_names[idx2])
# save_fig("iris-2d-unlabeled")
plt.show()


K = 3
y_pred = GaussianMixture(n_components=K, random_state=42).fit(X).predict(X)
mapping = np.array([2, 0, 1])
y_pred = np.array([mapping[cluster_id] for cluster_id in y_pred])

colors = sns.color_palette()[0:K]
markers = ("s", "x", "o", "^", "v")
fig, ax = plt.subplots()
for k in range(0, K):
    ax.plot(
        X[y_pred == k, idx1],
        X[y_pred == k, idx2],
        color=colors[k],
        marker=markers[k],
        linestyle="None",
        label="Cluster {}".format(k),
    )
ax.set(xlabel=iris.feature_names[idx1])
ax.set(ylabel=iris.feature_names[idx2])
plt.legend(loc="upper left", fontsize=12)
# save_fig("iris-2d-gmm")
plt.show()