Clustering Text Documents using K-Means in Scikit Learn

Clustering text documents is a typical issue in natural language processing (NLP). Based on their content, related documents are to be grouped. The k-means clustering technique is a well-liked solution to this issue. In this article, we’ll demonstrate how to cluster text documents using k-means using Scikit Learn.

K-means clustering algorithm

The k-means algorithm is a well-liked unsupervised learning algorithm that organizes data points into groups based on similarities. The algorithm operates by iteratively assigning each data point to its nearest cluster centroid and then recalculating the centroids based on the newly formed clusters.

Preprocessing

Preprocessing describes the procedures used to get data ready for machine learning or analysis. It frequently involves transforming, reformatting, and cleaning raw data and vectorization into a format appropriate for additional analysis or modeling.

Steps

1. Loading or preparing the dataset [dataset link: https://github.com/PawanKrGunjan/Natural-Language-Processing/blob/main/Sarcasm%20Detection/sarcasm.json]
2. Preprocessing of text in case the text is loaded instead of manually adding it to the code
3. Vectorizing the text using TfidfVectorizer
4. Reduce the dimension using PCA
5. Clustering the documents
6. Plot the cluster using matplotlib

Python3

# import the necessary libraries
import json
import numpy as np
import pandas as pd
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.decomposition import PCA
from sklearn.cluster import KMeans
import matplotlib.pyplot as plt

# Dataset link:
# https://github.com/PawanKrGunjan/Natural-Language-Processing/blob/main/Sarcasm%20Detection/sarcasm.json
df=pd.read_json('sarcasm.json')

# Extract the sentence only
sentence = df.headline

# create vectorizer
vectorizer = TfidfVectorizer(stop_words='english')

# vectorizer the text documents
vectorized_documents = vectorizer.fit_transform(sentence)

# reduce the dimensionality of the data using PCA
pca = PCA(n_components=2)
reduced_data = pca.fit_transform(vectorized_documents.toarray())


# cluster the documents using k-means
num_clusters = 2
kmeans = KMeans(n_clusters=num_clusters, n_init=5,
                max_iter=500, random_state=42)
kmeans.fit(vectorized_documents)


# create a dataframe to store the results
results = pd.DataFrame()
results['document'] = sentence
results['cluster'] = kmeans.labels_

# print the results
print(results.sample(5))

# plot the results
colors = ['red', 'green']
cluster = ['Not Sarcastic','Sarcastic']
for i in range(num_clusters):
    plt.scatter(reduced_data[kmeans.labels_ == i, 0],
                reduced_data[kmeans.labels_ == i, 1],
                s=10, color=colors[i],
                label=f' {cluster[i]}')
plt.legend()
plt.show()

Output:

                                                document  cluster
16263  study finds majority of u.s. currency has touc...        0
5318   an open and personal email to hillary clinton ...        0
12994        it's not just a muslim ban, it's much worse        0
5395   princeton students confront university preside...        0
24591     why getting married may help people drink less        0

Text clustering using KMeans