An Explorative Study on Extractive Text Summarization through k-means, LSA, and TextRank

Notably the difficult and exciting issue in the field of Natural Language Processing (NLP) is summarizing the text. Understanding the main objective of any type of document is crucial. Some of the applications of text summarization are media monitoring, social media, marketing, health care, literature, and books. Text summarization techniques are implemented using extractive summarization techniques in the health care domain in which it considers patient health history. To visualize a lengthy patient health history document quickly we use machine learning techniques like k-means, Text Rank, and Latent Semantic Analysis to comprehend and identify the sections that communicate important information to produce the summarized texts. These methods are evaluated using ROUGE-1, ROUGE-2, and ROUGE-N metrics to obtain the highest similarity of extracted text. k-means outperformed the considered approaches compared to Text Rank and Latent Semantic Analysis in summarizing the documents. k-Means was more efficient, where it achieved an average of 94.52% precision, 90.98% recall, and 91.25% F1-score.