A Comprehensive Review of Marathi Text Summarization Techniques

Aarya Shah, Dev Patel, Sagar Salvi, Minal Sonkar

Abstract: The summary of text is the part of Natural Language Processing (NLP), which deals with condensing long text into short, meaningful and remaining important pieces of text. Text summarization in regional languages is examined specifically, with focus on their linguistic and cultural problems along with text summarization techniques. Two prominent approaches are explored: extractive summarization, and abstractive summarization which filters important sentences or phrases from the original text, and generates new phrases that contain the essence of the content. Each algorithm is evaluated, used, and applied on regional language datasets with strengths and limitations of the techniques highlighted. Because they are simple and effective for preserving semantic integrity of the source text, extractive methods are more commonly applied. In contrast, the more natural and human-like summaries produced by the abstractive methods have two major drawbacks: semantic inaccuracies and inconsistent results, to name a few, and the difficult task of dealing with various linguistic structures. Finally, this paper highlights the importance of continued research and creating such models more reliably, so we may utilize them to support text summarization in regional languages.

Keywords: Natural Language Processing, Regional Language Text Summarization, Extractive Summarization, Abstractive Summarization, Term Frequency-Inverse Document Frequency (TF-IDF), TextRank Algorithm, Long Short Term Memory (LSTM)

How to Cite?: Aarya Shah, Dev Patel, Sagar Salvi, Minal Sonkar, "A Comprehensive Review of Marathi Text Summarization Techniques ", Volume 14 Issue 6, June 2025, International Journal of Science and Research (IJSR), Pages: 871-875, https://www.ijsr.net/getabstract.php?paperid=SR25604175059, DOI: https://dx.doi.org/10.21275/SR25604175059