Manuscript Number : SHISRRJ247256

A Stock Price Prediction Method Based Bi-LSTM and Improved Transformer

Authors(2) :-K. Naresh, A. V. Akshaya

Stock price prediction plays a crucial role in financial markets, aiding investors in making informed decisions. Traditional methods often struggle to capture the complex dynamics of stock prices due to their non-linear and volatile nature. In this study, we propose a novel approach that combines Bidirectional Long Short-Term Memory (Bi-LSTM) networks and Temporal Convolution Network (TCN) with an improved Transformer architecture to enhance stock price prediction accuracy. The proposed method leverages the temporal dependencies in stock price data through Bi-LSTM networks, which can effectively capture both past and future information simultaneously. Additionally, we introduce an improved Transformer architecture called Bi-LSTM-MTRAN-TCN that incorporates attention mechanisms to capture long-range dependencies and temporal patterns in the data in a better way. To evaluate the performance of our method, we conduct experiments on real-world stock price datasets. Comparative analyses are performed against baseline models, including traditional time series forecasting methods and single-model approaches. Our results demonstrate that the proposed method outperforms existing techniques in terms of prediction accuracy, particularly in capturing sudden market changes and complex patterns. Furthermore, we conduct sensitivity analyses to assess the robustness of our model to variations in input data and hyper parameters. The results indicate the effectiveness and stability of our approach across different market conditions and time periods.

Authors and Affiliations

K. Naresh
Assistant Professor, Department of MCA, Annamacharya Institute of Technology & Sciences, Tirupati, Andhra Pradesh, India
A. V. Akshaya
Post Graduate, Department of MCA, Annamacharya Institute of Technology & Sciences, Tirupati, Andhra Pradesh, India

Stock Price, Deep Learning Algorithms, Long Term Predictions, Bidirectional Long Short-Term Memory (Bi-LSTM), Temporal Convolution Network (TCN), Improved Transformer Models.

  1. Z. Hu, W. Liu, J. Bian, X. Liu, and T.-Y. Liu, "Listening to chaotic whis- pers: A deep learning framework for news-oriented stock trend prediction, in Proc. 11th ACM Int. Conf. Web Search Data Mining, Marina Del Rey, CA, USA, Feb. 2018, pp. 261–269.
  2. C. Xiao, W. Xia, and J. Jiang, "Stock price forecast based on combined model of ARI-MA-LS-SVM," Neural Comput. Appl., vol. 32, no. 10, pp. 5379–5388, May 2020
  3. W. Lu, J. Li, J. Wang, and L. Qin, "A CNN-BiLSTM-AM method for stock price prediction," Neural Comput. Appl., vol. 33, no. 10, pp. 4741–4753, May 2021.
  4. Y. Gao, R. Wang, and E. Zhou, "Stock prediction based on optimized LSTM and GRU models," Sci. Program., vol. 2021, pp. 1–8, Sep. 2021
  5. Y. Zhiyong, Y. Yuxi, and Z. Yu, "Application of BiLSTM-SA-TCN time series model in stock forecasting," J. Nanjing Inf. Eng. Univ., pp. 1–12, Apr. 2023
  6. Q. Yang and C. Wang, "Research on global stock index prediction based on deep learning LSTM neural network," Stat. Res., vol. 36, no. 3, pp. 65–77, 2019.

Publication Details

Published in : Volume 7 | Issue 2 | March-April 2024
Date of Publication : 2024-04-30
License:  This work is licensed under a Creative Commons Attribution 4.0 International License.
Page(s) : 252-260
Manuscript Number : SHISRRJ247256
Publisher : Shauryam Research Institute

ISSN : 2581-6306

Cite This Article :

K. Naresh, A. V. Akshaya, "A Stock Price Prediction Method Based Bi-LSTM and Improved Transformer", Shodhshauryam, International Scientific Refereed Research Journal (SHISRRJ), ISSN : 2581-6306, Volume 7, Issue 2, pp.252-260, March-April.2024
URL : https://shisrrj.com/SHISRRJ247256

Article Preview