Qolbi, Kintan Marezka Nurul (2025) Implementasi Text Mining Metode Ekstraksi Textrank dan Naive Bayes Classifier untuk Klasifikasi Jurnal Informatika Berdasarkan Abstrak Berbasis Website. Skripsi thesis, Universitas Jenderal Soedirman.

	PDF (Cover) COVER-Kintan Marezka-H1D021041-Tugas Akhir-2025.pdf Download (121kB)
	PDF (Legalitas) LEGALITAS-Kintan Marezka-H1D021041-Tugas Akhir-2025.pdf Restricted to Repository staff only Download (292kB)
	PDF (Abstrak) ABSTRAK-Kintan Marezka-H1D021041-Tugas Akhir-2025.pdf Download (237kB)
	PDF (BabI) BAB I-Kintan Marezka-H1D021041-Tugas Akhir-2025.pdf Restricted to Repository staff only until 5 November 2026. Download (243kB)
	PDF (BabII) BAB II-Kintan Marezka-H1D021041-Tugas Akhir-2025.pdf Restricted to Repository staff only until 5 November 2026. Download (425kB)
	PDF (BabIII) BAB III-Kintan Marezka-H1D021041-Tugas Akhir-2025.pdf Restricted to Repository staff only until 5 November 2026. Download (368kB)
	PDF (BabIV) BAB IV-Kintan Marezka-H1D021041-Tugas Akhir-2025.pdf Restricted to Repository staff only Download (803kB)
	PDF (BabV) BAB V-Kintan Marezka-H1D021041-Tugas Akhir-2025.pdf Download (218kB)
	PDF (DaftarPustaka) DAFTAR PUSTAKA-Kintan Marezka-H1D021041-Tugas Akhir-2025.pdf Download (197kB)

Abstract

The rapid advancement of information technology has led to a significant increase in scientific publications, creating challenges in efficiently and accurately retrieving relevant information. This study aims to develop a web-based automatic journal classification system using the Textrank and Naïve Bayes Classifier methods. The Textrank method is used to extract important keywords from journal abstracts, while the Naive Bayes Classifier is employed to categorize journal abstracts, while the Naive Bayes Classifier is employed to categorize journals into six informatics-related categories: Artificial Intelligence, Software Engineering, Computer Networks, Information Security, Information Systems, and Data Science. The dataset consists of 180 journals for training and 30 journals for testing. The classification process involves text preprocessing, keyword extraction (up to 25 keywords), and probability calculation using Bayes theorem. The system was tested using the black box method and evaluated through a confusion matrix. The results show that the combination of Textrank and Naive Bayes methods can classify journals with an accuracy of 93%, precision of 85%, and recall of 80%. This system is expected to assist users in finding relevant journals more quickly and efficiently.

Actions (login required)

View Item