A Comparative Survey of Text Summarization Techniques

Text summarization holds significance in the realm of natural language processing as it expedites the extraction of crucial information from extensive textual content. The paper presents an overview of six prevalent techniques for text summarization: TextRank, which identifies key phrases and sentences based on Google's PageRank algorithm; ChatGPT, blending extractive and abstractive methods; alongside BERT, LSA, and BART. The assessment of these strategies involves their categorization into two aspects: (1) the level of summarization achieved and (2) the degree of readability attained. Consequently, a survey was conducted involving 15 participants, employing a questionnaire to gauge the clarity and informative nature of the generated summaries. Complementing this qualitative approach, quantitative evaluations employing ROUGE scores were employed. The outcomes of the survey offer a comprehensive comparative analysis of these techniques, revealing their efficacy in generating accurate and comprehensible summaries. These empirical data assist in selecting the optimal technique for a given task and offers insights for future investigations. © 2023, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.