Employing a Novel Metaheuristic Algorithm to Optimize an LSTM Model: A Case Study of Stock Market Prediction
Research output: Contributions to collected editions/works › Article in conference proceedings › Research › peer-review
Standard
Advances in Computing and Data Sciences: 7th International Conference, ICACDS 2023, Kolkata, India, April 27–28, 2023, Revised Selected Papers. ed. / Mayank Singh; Vipin Tyagi; P.K. Gupta; Jan Flusser; Tuncer Ören. Cham: Springer International Publishing AG, 2023. p. 478-489 (Communications in Computer and Information Science; Vol. 1848).
Research output: Contributions to collected editions/works › Article in conference proceedings › Research › peer-review
Harvard
APA
Vancouver
Bibtex
}
RIS
TY - CHAP
T1 - Employing a Novel Metaheuristic Algorithm to Optimize an LSTM Model
T2 - International Conference on Advances in Computing and Data Sciences
AU - Dastgerdi, Amin Karimi
AU - Mercorelli, Paolo
N1 - Conference code: 7
PY - 2023/1/1
Y1 - 2023/1/1
N2 - It has been long since researchers as well as investors and stakeholders who are actively pursuing financial markets are trying to analyze stock price movements and predict its trend more accurately. To minimize the forecasting risk and make the most profit, several methods have been used among which Deep Learning was at the center of attention in recent years. Deep learning techniques include analyses of historical data and recognizing patterns that can assist scientists to make a more precise prediction. This paper focuses on the application of a optimization approach called Neural Network Algorithm to optimize Long short-term Memory for the prediction of financial time series. The findings reveal that the utilization of an optimization technique such as Neural Network Algorithm to optimize Long Short-term Memory neural networks results in a notable improvement of 40%, 65%, 4%, and 85% in the MAPE, Theil U, R, and RMSE metrics, respectively. Consequently, this leads to even more accurate results and more precise predictions.
AB - It has been long since researchers as well as investors and stakeholders who are actively pursuing financial markets are trying to analyze stock price movements and predict its trend more accurately. To minimize the forecasting risk and make the most profit, several methods have been used among which Deep Learning was at the center of attention in recent years. Deep learning techniques include analyses of historical data and recognizing patterns that can assist scientists to make a more precise prediction. This paper focuses on the application of a optimization approach called Neural Network Algorithm to optimize Long short-term Memory for the prediction of financial time series. The findings reveal that the utilization of an optimization technique such as Neural Network Algorithm to optimize Long Short-term Memory neural networks results in a notable improvement of 40%, 65%, 4%, and 85% in the MAPE, Theil U, R, and RMSE metrics, respectively. Consequently, this leads to even more accurate results and more precise predictions.
KW - Engineering
KW - Deep Learning
KW - Time Series Forecasting
KW - LSTM
KW - Optimization
KW - Metaheuristic Algorithms
UR - http://www.scopus.com/inward/record.url?scp=85172195374&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-37940-6_39
DO - 10.1007/978-3-031-37940-6_39
M3 - Article in conference proceedings
SN - 978-3-031-37939-0
T3 - Communications in Computer and Information Science
SP - 478
EP - 489
BT - Advances in Computing and Data Sciences
A2 - Singh, Mayank
A2 - Tyagi, Vipin
A2 - Gupta, P.K.
A2 - Flusser, Jan
A2 - Ören, Tuncer
PB - Springer International Publishing AG
CY - Cham
Y2 - 27 April 2023 through 28 April 2023
ER -