Authors:
B. Gunapriya, M. Arunadevi Thirumalraj, B. Manjunatha, S. Gopikha, Robert Balku

Addresses:
Department of Electrical and Electronics Engineering, New Horizon College of Engineering, Bengaluru, Karnataka, India. Department of Computer Science and Engineering, Karunya Institute of Technology and Science, Coimbatore, Tamil Nadu, India. Department of Computer Science and Business Management, Saranathan College of Engineering, Tiruchirappalli, Tamil Nadu, India. Department of Mechanical Engineering, New Horizon College of Engineering, Bengaluru, Karnataka, India. Department of Information Technology, St. Joseph's College of Engineering, Chennai, Tamil Nadu, India. Department of Business Administration, Budapest Metropolitan University, Budapest, Hungary.

Abstract:

The complex, dynamic, and non-linear nature of stock group value projection has historically made it both an enticing and demanding task for shareholders. An integral part of investing in stocks is accurately forecasting their prices. Stock price data is notoriously difficult to forecast with any degree of accuracy due to its high frequency, nonlinearity, and lengthy memory. This article focuses on the market predictions of GOF groups. Experimental evaluations were conducted using datasets such as the Tehran Stock Exchange and the Chinese Stock Index, specifically the CSI300 index. For each category, we combed through a decade-long database for relevant information. A1,2,5,10-,15-,20-, and 30-day time horizons are included in the value forecasts. A 1-D convolutional neural network is used to extract features, which are then passed to classifiers. A new predictive model is introduced here that effectively extracts nodes' trust levels from their spatial and temporal performance metrics by combining the power of VARMA, GRU, and LSTM neural networks. Doctor and patient optimisation (DPO) enhances classification accuracy by selecting the LSTM weight appropriately. When compared to LSTM, RNN, and CNN, the proposed model consistently outperforms them all in predicting stock prices. 

Keywords: Vector Autoregressive; Moving Average; Long Memory; Gated Recurrent Unit; Doctor and Patient Optimisation; Stock Prediction; Non-Linear Nature; LSTM Neural Networks.

Received on: 04/12/2024, Revised on: 27/02/2025, Accepted on: 09/04/2025, Published on: 07/09/2025

DOI: 10.69888/FTSCS.2025.000478

FMDB Transactions on Sustainable Computing Systems, 2025 Vol. 3 No. 3, Pages: 203-214

• Views : 79
• Downloads : 18

Download PDF