Journal Technology Information and Data Analytic
Vol 2 No 1 (2025): Journal Technology Information and Data Analytic (TIFDA)

Implementasi Natural Language Processing (NLP) dalam Pengembangan Aplikasi Chatbot untuk Pembelajaran Teks Islam Klasik di Pesantren El-Huda El-Islamy

Sofyan, Yan (Unknown)
Arroyan, Alwi Fachri Ibnu (Unknown)



Article Info

Publish Date
20 Jun 2025

Abstract

El-Huda El-Islamy Islamic Boarding School is an Islamic educational institution that aims to mould students into competent and noble religious preachers. This institution focuses on learning and deep understanding of Islamic teachings, including the study of classical books such as the Yellow Book of Fathul Qorib. This research aims to develop and evaluate an Android-based chatbot application that uses the BERT model to help the learning process of the Yellow Book of Fathul Qorib at El-Huda Islamic Boarding School. In this research, the BERT model is integrated into the chatbot application to understand and respond to user questions appropriately. Testing was done by asking 25 questions to the chatbot, which successfully answered 11 questions with a success rate of 44%. Evaluation of the model performance using confusion matrix showed that the chatbot had 90% accuracy, 87% precision, 87% recall, and 86% F1-score. These results show that the chatbot has not been able to provide relevant and accurate responses, and recognise most of the questions asked. This research concludes that this chatbot application is not yet an effective tool to support the learning process at El-Huda Islamic Boarding School

Copyrights © 2025






Journal Info

Abbrev

tifda

Publisher

Subject

Agriculture, Biological Sciences & Forestry Computer Science & IT Decision Sciences, Operations Research & Management Engineering Library & Information Science Other

Description

Journal of Technology Information and Data Analytic is a scientific journal managed by the Faculty of Engineering, Darma Persada University. TIFDA is an open access journal that provides free access to the full text of all published articles without charging access fees from readers or their ...