Comparative Analysis of Clustering Approaches for Text Summarization of Covid's News

The outbreak of COVID-19 resulted in a massive surge of online news articles, updates, and reports, creating an overwhelming volume of information for readers to process. To help users quickly grasp essential content, this project focuses on the comparative analysis of clustering-based approaches for automatic text summarization of COVID-19 news. The goal is to identify the most efficient technique for extracting concise, relevant summaries from large collections of articles. This study evaluates different clustering algorithms such as K-Means, Hierarchical Clustering, and DBSCAN, considering their ability to group similar information and highlight key topics. The preprocessing steps include tokenization, stop word removal, and vectorization using TF-IDF to improve clustering performance. The generated summaries are assessed based on coherence, content coverage, redundancy reduction, and computational efficiency. Results indicate that while K-Means offers speed and scalability, Hierarchical Clustering delivers more structured summaries, and DBSCAN effectively filters noise. This analysis not only highlights the strengths and trade-offs of each method but also offers guidance for building more accurate and responsive summarization systems, especially during high-impact global events like pandemics.