A Web-Based Document Summarizing and Topic Prediction System Using Natural Language Processing

Abstract: The need to read through large volumes of information or other bodies of text and produce a summarized version of the documents is usually a daunting task for many. For instance, researchers often times find themselves reading through lines, paragraphs, pages or chapters of documents before finally coming across the key points which will help them with their research work. Most existing systems generate the summary only in text and a specific language. Therefore, this study provides a web-based document summarizing and topic prediction system with Text-To-Speech and auto-language detection features. The system provides two types of two types of summaries; Extractive and abstractive summaries. The abstractive summary feature makes use of the GPT-3 API while the extractive summary and topic prediction make use of sentence scoring. The language detection model is trained with language data found on Kaggle. The text-to-speech feature is implemented using a JavaScript library. The system is able to detect the topic of the body of text and displays it to the user, and identifies the language of the text entered while also providing an appropriate summary. Users are able to play audio summaries and download them as a word document. Experimental results of the system have found that both the extractive and abstractive summary features are able to capture the main points of any given body of text and present summaries that convey information about the entire body of text, leveraging on Artificial Intelligence (AI) with a focus on Natural Language Processing (NLP). Keywords: Artificial Intelligence, Natural Language Processing, Topic Prediction, Language Detection, Text-to-Speech. Title: A Web-Based Document Summarizing and Topic Prediction System Using Natural Language Processing Author: Taiwo Adigun, Einstein Ebereonwu International Journal of Recent Research in Mathematics Computer Science and Information Technology ISSN 2350-1022 Vol. 11, Issue 2, October 2024 - March 2025 Page No: 74-83 Paper Publications Website: www.paperpublications.org Published Date: 13-February-2025 DOI: https://doi.org/10.5281/zenodo.14863280 Paper Download Link (Source) https://www.paperpublications.org/upload/book/A%20Web-Based%20Document%20Summarizing-13022025-2.pdf