Predicting the optimum number of clusters from a dataset using Python

In this tutorial, we are exploring unsupervised machine learning using Python. We will predict the optimum number of clusters from iris dataset and visualize it. This tutorial will walk through some of the basics of K-Means Clustering.

Exploring unsupervised machine learning with the iris dataset

program code:

importing all the required libraries to the python notebook

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
%matplotlib inline

Loading the iris dataset

iris = datasets.load_iris()
iris_df = pd.DataFrame(, columns = iris.feature_names)
#Displaying the whole dataset
# Displaying the first 5 rows

Finding the optimum number of clusters for k-means classification and also showing how to determine the value of K

x = iris_df.iloc[:, [0, 1, 2, 3]].values

from sklearn.cluster import KMeans
wcss = []

for i in range(1, 11):
    kmeans = KMeans(n_clusters = i, init = 'k-means++', 
                    max_iter = 300, n_init = 10, random_state = 0)
# Plotting the results onto a line graph, 
# `allowing us to observe 'The elbow'
plt.plot(range(1, 11), wcss)
plt.title('The elbow method')
plt.xlabel('Number of clusters')
plt.ylabel('WCSS') # Within cluster sum of squares

You can perceive any reason why it is known as ‘The elbow technique’ from the above graph, the optimum clusters are the place the elbow happens. This is the point at which the WCSS(within-cluster sum of squares) doesn’t reduce essentially with each iteration.

From this, we choose the number of clusters as ** ‘3**’.

Applying k means to the dataset / Creating the k means classifier.

kmeans = KMeans(n_clusters = 3, init = 'k-means++',
                max_iter = 300, n_init = 10, random_state = 0)
y_kmeans = kmeans.fit_predict(x)

Plotting the centroids of the clusters

plt.scatter(kmeans.cluster_centers_[:, 0], kmeans.cluster_centers_[:,1], 
            s = 100, c = 'yellow', label = 'Centroids')


After executing all the above-given codes, The final result will be plotted centroids of the cluster in the graph. It shows the predicted optimum number of clusters from the iris dataset.


Leave a Reply

Your email address will not be published. Required fields are marked *