TY - GEN
T1 - Stock Transaction Strategy Using Deep Learning and Support-Resistance Level Methods
AU - Haidar, Ziyad
AU - Sarno, Riyanarto
AU - Anggraini, Ratih Nur Esti
AU - Haryono, Agus Tri
AU - Septiyanto, Abdullah Faqih
AU - Lee, Sang Seok
N1 - Publisher Copyright:
© 2025 IEEE.
PY - 2025
Y1 - 2025
N2 - The right decision on stock trading is important because it can impact stock gain or loss. The support-resistance level method is used to predict signal stock trading with good results. The deep learning method is already used for stock price forecasting, but stock return performance has not yet been compared. This study proposed comparing stock return from support-resistance level and deep learning method. The deep learning method is represented by Gated Recurrent Units (GRU), Long Short-Term Memory (LSTM), and a combination of the two models, each with two stacked layers: GRU-LSTM and LSTM-GRU. The backtesting method compares the best deep learning and a support-resistance level on stock return. Backtesting using stock trading signal by extracting result from support-resistance level and deep learning methods in 1,000 days. As the best performance on deep learning, the GRU model produces an average stock return of 256.33%.
AB - The right decision on stock trading is important because it can impact stock gain or loss. The support-resistance level method is used to predict signal stock trading with good results. The deep learning method is already used for stock price forecasting, but stock return performance has not yet been compared. This study proposed comparing stock return from support-resistance level and deep learning method. The deep learning method is represented by Gated Recurrent Units (GRU), Long Short-Term Memory (LSTM), and a combination of the two models, each with two stacked layers: GRU-LSTM and LSTM-GRU. The backtesting method compares the best deep learning and a support-resistance level on stock return. Backtesting using stock trading signal by extracting result from support-resistance level and deep learning methods in 1,000 days. As the best performance on deep learning, the GRU model produces an average stock return of 256.33%.
KW - Deep Learning
KW - GRU
KW - LSTM
KW - Stock Trading
KW - Support-Resistance Level
UR - https://www.scopus.com/pages/publications/105025479085
U2 - 10.1109/AIMS66189.2025.11229724
DO - 10.1109/AIMS66189.2025.11229724
M3 - Conference contribution
AN - SCOPUS:105025479085
T3 - 2025 IEEE International Conference on Artificial Intelligence and Mechatronics Systems, AIMS 2025
BT - 2025 IEEE International Conference on Artificial Intelligence and Mechatronics Systems, AIMS 2025
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 3rd IEEE International Conference on Artificial Intelligence and Mechatronics Systems, AIMS 2025
Y2 - 24 May 2025 through 25 May 2025
ER -