Text Clustering on News Dataset

In this notebook we use NLP methods and kernel PCA to build a clustering model

Dataset on kaggle

Contents

Data Cleaning

• Column selection
• Null values
• Duplicate values

Data Preprocessing

• Lowercasing and removing newline, tabular spaces, numbers and punctuations
• Stopwords removal
• Lemmatization
• Label encoding
• TF-IDF vectorization

Kernel PCA

In this part we apply kernel PCA to reduce dimensions

K-Means Clustering

At first, we apply the elbow method to calculate the best K and then build a model

Silhouette Score

The score for this task is 0.81

Visualization

We finish the task with final analyses