Wavelet-Enhanced Deep Learning Ensemble for Accurate Stock Market Forecasting: A Case Study of Nifty 50 Index

Portfolio theory underpins portfolio management, a much-researched yet uncharted field. Stock market prediction is a challenging and essential endeavour in financial research, owing to the nonlinear, volatile, and stochastic characteristics of financial time series data. Conventional statistical tec...

Full description

Saved in:
Bibliographic Details
Main Authors: Priya Singh, Manoj Jha, Harshita Patel
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11005722/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849714488184406016
author Priya Singh
Manoj Jha
Harshita Patel
author_facet Priya Singh
Manoj Jha
Harshita Patel
author_sort Priya Singh
collection DOAJ
description Portfolio theory underpins portfolio management, a much-researched yet uncharted field. Stock market prediction is a challenging and essential endeavour in financial research, owing to the nonlinear, volatile, and stochastic characteristics of financial time series data. Conventional statistical techniques often fall short to encapsulate complex interdependencies, resulting in diminished predictive accuracy. This research proposes an ensemble model that integrates Long Short-Term Memory (LSTM), Convolutional Neural Networks (CNN), and Temporal Convolutional Networks (TCN) for effective stock market prediction. The Nifty 50 index dataset is utilized for the empirical evidences. Wavelet-based denoising is utilised as a preprocessing measure to mitigate the intrinsic noise in stock market data. The model&#x2019;s efficacy is assessed utilising error metrics, such as Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and R-squared (<inline-formula> <tex-math notation="LaTeX">$R^{2}$ </tex-math></inline-formula>). The five-fold cross-validation is utilized to establish the robustness of the models. Furthermore, we ascertain the statistical significance of performance enhancements by parametric t-tests, including normality assessments via the Shapiro-Wilk test. Moreover, current state-of-the-art models advocates in favour of proposed study.
format Article
id doaj-art-3f5d89a26c6f4fcca7502fadb1c812d9
institution DOAJ
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-3f5d89a26c6f4fcca7502fadb1c812d92025-08-20T03:13:42ZengIEEEIEEE Access2169-35362025-01-0113870368704710.1109/ACCESS.2025.356863411005722Wavelet-Enhanced Deep Learning Ensemble for Accurate Stock Market Forecasting: A Case Study of Nifty 50 IndexPriya Singh0https://orcid.org/0000-0002-0625-9617Manoj Jha1Harshita Patel2https://orcid.org/0000-0002-9381-9028School of Advanced Sciences, Vellore Institute of Technology, Vellore, IndiaDepartment of Mathematics, Bioinformatics and Computer Applications, Maulana Azad National Institute of Technology, Bhopal, Madhya Pradesh, IndiaSchool of Computer Science Engineering and Information Systems, Vellore Institute of Technology, Vellore, IndiaPortfolio theory underpins portfolio management, a much-researched yet uncharted field. Stock market prediction is a challenging and essential endeavour in financial research, owing to the nonlinear, volatile, and stochastic characteristics of financial time series data. Conventional statistical techniques often fall short to encapsulate complex interdependencies, resulting in diminished predictive accuracy. This research proposes an ensemble model that integrates Long Short-Term Memory (LSTM), Convolutional Neural Networks (CNN), and Temporal Convolutional Networks (TCN) for effective stock market prediction. The Nifty 50 index dataset is utilized for the empirical evidences. Wavelet-based denoising is utilised as a preprocessing measure to mitigate the intrinsic noise in stock market data. The model&#x2019;s efficacy is assessed utilising error metrics, such as Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and R-squared (<inline-formula> <tex-math notation="LaTeX">$R^{2}$ </tex-math></inline-formula>). The five-fold cross-validation is utilized to establish the robustness of the models. Furthermore, we ascertain the statistical significance of performance enhancements by parametric t-tests, including normality assessments via the Shapiro-Wilk test. Moreover, current state-of-the-art models advocates in favour of proposed study.https://ieeexplore.ieee.org/document/11005722/Ensemble learningstock marketrandom forestLSTMTCNCNN
spellingShingle Priya Singh
Manoj Jha
Harshita Patel
Wavelet-Enhanced Deep Learning Ensemble for Accurate Stock Market Forecasting: A Case Study of Nifty 50 Index
IEEE Access
Ensemble learning
stock market
random forest
LSTM
TCN
CNN
title Wavelet-Enhanced Deep Learning Ensemble for Accurate Stock Market Forecasting: A Case Study of Nifty 50 Index
title_full Wavelet-Enhanced Deep Learning Ensemble for Accurate Stock Market Forecasting: A Case Study of Nifty 50 Index
title_fullStr Wavelet-Enhanced Deep Learning Ensemble for Accurate Stock Market Forecasting: A Case Study of Nifty 50 Index
title_full_unstemmed Wavelet-Enhanced Deep Learning Ensemble for Accurate Stock Market Forecasting: A Case Study of Nifty 50 Index
title_short Wavelet-Enhanced Deep Learning Ensemble for Accurate Stock Market Forecasting: A Case Study of Nifty 50 Index
title_sort wavelet enhanced deep learning ensemble for accurate stock market forecasting a case study of nifty 50 index
topic Ensemble learning
stock market
random forest
LSTM
TCN
CNN
url https://ieeexplore.ieee.org/document/11005722/
work_keys_str_mv AT priyasingh waveletenhanceddeeplearningensembleforaccuratestockmarketforecastingacasestudyofnifty50index
AT manojjha waveletenhanceddeeplearningensembleforaccuratestockmarketforecastingacasestudyofnifty50index
AT harshitapatel waveletenhanceddeeplearningensembleforaccuratestockmarketforecastingacasestudyofnifty50index