Question answering system zakat dengan metode Long Short-Term Memory (LSTM)

Tanuwijaya, Moch Apip and Jumadi, Jumadi and Nurlatifah, Eva (2025) Question answering system zakat dengan metode Long Short-Term Memory (LSTM). Bulletin of Computer Science Research, 5 (5). pp. 929-938. ISSN 2774-3659

Preview

Text 728-Article Text-3086-1-10-20250809.pdf Download (848kB) | Preview

Abstract

INDONESIA: Zakat merupakan salah satu pilar penting dalam sistem keuangan Islam yang berfungsi sebagai instrumen redistribusi kekayaan. Namun, belum tersedia sistem Question Answering System (QAS) berbahasa Indonesia yang mampu memberikan jawaban otomatis dan kontekstual terkait zakat. Penelitian ini bertujuan mengembangkan sistem QAS zakat berbasis Long Short-Term Memory (LSTM) yang terintegrasi dengan platform Telegram. Dataset dikumpulkan dari buku panduan resmi BAZNAS dan diproses melalui tokenisasi, padding, dan label encoding. Arsitektur model terdiri dari layer embedding, dua layer LSTM bertingkat (dengan return_sequences, dropout, dan recurrent dropout), serta dua layer dense bertingkat (200 dan 100 unit) dengan dropout tambahan sebelum output softmax. Model dilatih menggunakan Adam optimizer (learning rate 0.003), batch size 24, dan 100 epoch. Evaluasi dilakukan menggunakan confusion matrix dengan hasil akurasi validasi sebesar 93%, precision 0.94, recall 0.93, dan F1-score 0.92 (weighted average). Sistem diintegrasikan ke Telegram Bot API dengan respons di bawah dua detik dan mampu menangani ratusan label kelas secara stabil. Sistem ini menunjukkan potensi dalam mendukung edukasi zakat digital dan dapat dikembangkan lebih lanjut dalam ekosistem Islamic Finance Technology dan Digital Religious Education. ENGLISH: Zakat is a fundamental pillar of Islamic finance that serves as a mechanism for wealth redistribution. However, there is currently no Indonesian-language Question Answering System (QAS) capable of automatically and contextually responding to zakatrelated queries. This study aims to develop a zakat-focused QAS using a Long Short-Term Memory (LSTM) model integrated into the Telegram platform. The dataset was compiled from the official BAZNAS zakat guidebook and processed through tokenization, padding, and label encoding. The model architecture consists of an embedding layer, two stacked LSTM layers (with return sequences, dropout, and recurrent dropout), followed by two dense layers (200 and 100 units) with additional dropout layers before the softmax output. The model was trained using the Adam optimizer (learning rate 0.003), a batch size of 24, and 100 epochs. Evaluation was conducted using a confusion matrix, resulting in a validation accuracy of 93%, with a precision of 0.94, recall of 0.93, and F1-score of 0.92 (weighted average). The system was deployed via the Telegram Bot API and demonstrated response times under two seconds, with stable performance across hundreds of question labels. This work contributes to the advancement of digital zakat education and presents a scalable solution that can be further extended within the ecosystem of Islamic Finance Technology and Digital Religious Education.

Item Type:	Article
Uncontrolled Keywords:	Question Answering System; LSTM; NLP; Zakat; Telegram Bot API; Islamic Finance Technology; Digital Religious Education
Subjects:	Data Processing, Computer Science Data Processing, Computer Science > Computer and Human Special Computer Methods Special Computer Methods > Artificial Intelligence Engineering
Divisions:	Fakultas Sains dan Teknologi > Program Studi Teknik Informatika
Depositing User:	Moch Apip Tanuwijaya
Date Deposited:	07 Nov 2025 23:04
Last Modified:	07 Nov 2025 23:04
URI:	https://digilib.uinsgd.ac.id/id/eprint/125234

Available Versions of this Item

• Question answering system zakat dengan metode Long Short-Term Memory (LSTM). (deposited 07 Nov 2025 23:04) [Currently Displayed]

Actions (login required)

View Item