Comparative Analysis of TextRank and Latent Semantic Analysis Algorithms for Extractive News Summarization

News summarization is an essential process that condenses news stories or collections of stories while retaining key details and main ideas. It offers readers a concise overview of significant events and essential takeaways from news stories, allowing them to quickly grasp the necessary information without reading the entire article. This paper presents a comparative analysis of two widely used algorithms, TextRank, a graph-based algorithm based on the PageRank algorithm, and Latent Semantic Analysis (LSA), a statistical method that captures latent semantic relationships using singular value decomposition for news summarization. The experiment results obtained using established evaluation metrics such as ROUGE-N and ROUGE-L suggest that the TextRank algorithm performs better than the LSA algorithm for all news categories and consistently demonstrates higher Precision, Recall, and F-Score values. The research aims to contribute to a deeper understanding of the strengths and weaknesses of TextRank and LSA in summarizing news by incorporating five news categories (Business, Politics, Sports, Entertainment, and Technology) and using a significantly large dataset containing 2225 news articles and their summaries. This knowledge will empower researchers and practitioners to make informed decisions when selecting an algorithm for summarization tasks based on their requirements and constraints.