Enhancing Extractive Text Summarization through Firefly Algorithm Optimization

This paper presents a novel approach to extractive text summarization using the Firefly algorithm, a bio-inspired optimization technique. The method involves evaluating candidate summaries based on four factors: Topic Relation Factor (TRF), Cohesion Factor (CF), Readability Factor (RF), and Content Coverage Factor (C). Each factor contributes to the overall fitness function, which guides the Firefly algorithm in selecting the most relevant sentences from a document. By representing sentences as binary vectors and using the TF-IDF method for similarity calculations, the algorithm iteratively updates fireflies, optimizing the summary’s quality. Experimental results, evaluated using ROUGE metrics, demonstrate the effectiveness of this approach in generating coherent and informative summaries, outperforming traditional methods.