Title: Optimised hybrid CNN-LSTM model for stock price prediction
Authors: Deepti Patnaik; N.V. Jagannadha Rao; Brajabandhu Padhiari; Srikanta Patnaik
Addresses: School of Management Studies, Gandhi Institute of Engineering and Technology University (GIETU), Gunupur, Rayagada, Odisha, 765022, India ' School of Management Studies, Gandhi Institute of Engineering and Technology University (GIETU), Gunupur, Rayagada, Odisha, 765022, India ' Interscience Institute of Management and Technology (IIMT) Bhubaneswar, Kantabada, Bhubaneswar, Khurda (Dt), Odisha 752024, India ' Interscience Institute of Management and Technology (IIMT) Bhubaneswar, Kantabada, Bhubaneswar, Khurda (Dt), Odisha 752024, India
Abstract: This article emphasises the precise forecasting of stock prices using a hybrid deep learning model that is a convolutional neural network - long short-term memory network and the parameters are optimised by enhanced grey wolf optimiser (GWO). With the availability of huge data in the present scenario, deep learning models outperform better than all other models. Again, to avoid the slow convergence rate and stagnation of local optima, an enhanced GWO algorithm is used. Stock prices of more than 12 years of various challenging stock exchanges such as: Standard & Poor 500, NIFTY 50, Nikkei 225, Dow Jones are used here for analysis purposes. Performance parameters used are root mean square error, mean absolute percentage error and mean absolute error. The proposed hybrid model is also compared with state-of-art models and it is found that this proposed model performs better than the existing models.
Keywords: forecasting; convolutional neural network; CNN; long short-term memory; LSTM; hybrid model; evolutionary computation; enhanced grey wolf optimisation; GWO; root mean square error; RMSE; mean absolute percentage error; MAPE; mean absolute error; MAE.
DOI: 10.1504/IJMDM.2024.139387
International Journal of Management and Decision Making, 2024 Vol.23 No.4, pp.438 - 460
Received: 17 Oct 2022
Accepted: 17 Dec 2022
Published online: 02 Jul 2024 *