Implementation of Natural Language Processing for Text Summarization in a Thesis Using The Textrank and Lexrank Algorithms

Abstract

Mercu Buana University is one of the higher education institutions in Indonesia, where, during their studies, students will face the final phase of their education, which is the thesis defense. In preparing their thesis, students require references from online repositories and university libraries, so the implementation of Natural Language Processing (NLP) for text summarization in the context of student thesis writing can help them complete their tasks. With the increasing number of references that need to be read, many students face difficulties and tend to be reluctant to read through the entire text. This can hinder their understanding and the quality of the thesis they produce. Therefore, this study aims to develop an automatic summarization system that can assist students in understanding the core of the existing references without the need to read the entire text. The methods used in this study include data collection from various types of academic references, text processing using Natural Language Processingtechniques such as tokenization, stopword removal, and syntactic analysis, as well as the application of extraction-based summarization algorithms using TextRank and LexRank. LexRank is a graph-based summarization algorithm that identifies important sentences in the text and selects the most relevant ones to include in the summary. The results of this system are expected to provide accurate and informative summaries, making it easier for students to utilize available resources. The system evaluation is conducted by involving students as end users to assess the effectiveness and satisfaction with the generated summaries. The findings of this study are expected to contribute to improving students' academic capabilities and encourage them to be more active in research and thesis writing.