Article Per Year (5 Year)
p-Index From 2021 - 2026
0.23
P-Index
This Author published in this journals
All Journal Kurawal - Jurnal Teknologi, Informasi dan Industri
Kamal, Muhamad Ramadhan
Unknown Affiliation
Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Industrial & Manufacturing Engineering Other

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Analisis Komparatif Kinerja LSTM, BiLSTM, dan LSTM-AM dalam Prediksi Harga Saham Syariah Taufik, Rahman; Kamal, Muhamad Ramadhan; Sholehurrahman, Ridho; ., Tristiyanto; Zen, Bita Parga
Kurawal - Jurnal Teknologi, Informasi dan Industri Vol 9 No 1 (2026): Vol 9 No 1 (2026): Jurnal Kurawal Volume 9, Nomor 1, Maret 2026
Publisher : Universitas Ma Chung

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33479/kurawal.v9i1.1470

Abstract

The continuously evolving Sharia stock market necessitates more accurate price modeling due to high volatility, regime shifts, and outliers that frequently disrupt investment decision-making processes. This study aims to comparatively evaluate the performance of three deep learning algorithms—LSTM (Long Short-Term Memory), BiLSTM (Bidirectional LSTM), and LSTM-AM (LSTM with Attention Mechanism)—in predicting Sharia stock prices. The research method utilizes daily closing price data from five Indonesian Sharia-compliant issuers (ANTM, ERAA, KLBF, SMGR, and WIKA) spanning the period of December 2016 to December 2021. The data underwent preprocessing using Robust Scaling and was structured into time series with a 60-day window. Model evaluation was conducted via window-based cross-validation with the results show that BiLSTM delivered the best performance with an average MAPE of 9.41% and RMSE of 249.956. This performance was followed by LSTM (MAPE 11.87%) and LSTM-AM as the lowest (MAPE 19.58%). These findings provide a clear understanding of the effectiveness of each architecture in predicting dynamic Sharia stock prices, indicating that increasing model complexity (LSTM-AM) does not always guarantee better accuracy in this domain. Consequently, BiLSTM can be considered the superior model for more stable prediction.
Co-Authors ., Tristiyanto Bita Parga Zen Sholehurrahman, Ridho Taufik, Rahman