Author
Listed:
- Shaid Hasan
(Department of Business Analytics, Trine University, Angola, Indiana, USA)
- Ismoth Zerine
(College of Graduate and Professional Studies, Trine University, Angola, Indiana, USA)
- Md Mainul Islam
(College of Graduate and Professional Studies, Trine University, Angola, Indiana, USA)
- Adib Hossain
(Department of Business Analytics, Trine University, Angola, Indiana, USA)
- Khandaker Ataur Rahman
(Department of Business Analytics, Trine University, Angola, Indiana, USA)
- Zulkernain Doha
(Faculty of Business and Technology, Grand Canyon University, USA)
Abstract
The stock markets are too complex to predict, despite their non-linearity, volatility, and multifactorial nature. Current econometric and deep learning models often do not take into account the macroeconomic context, which restricts predictive accuracy at times of financial volatility. This paper attempts to fill this gap by proposing a Hybrid Deep Learning Model (HDLM), a machine that combines Long Short-Term Memory (LSTM) networks, Convolutional Neural Networks (CNNs), and an attention mechanism that can jointly include both the dynamic behavior of the market and the economic links in the trends of the U.S. market. The analysis is based on a strictly tested data set of 20102023, which was retrieved from the Federal Reserve Economic Data (FRED), Yahoo Finance, and the Bureau of Economic Analysis. These include all significant equity indexes such as the S&P 500, NASDAQ Composite, and Dow Jones Industrial Average, prime macroeconomic indicators such as the GDP growth, inflation rates, interest rates, and unemployment levels. Correlation analysis, multiple regression, Granger causality tests, Johansen cointegration procedures, and volatility modeling using ARCH GARCH models are the statistical methods used. The findings show that GDP growth has a statistically significant positive impact on market returns ( = 0.58, p = 0.001), though inflation ( = -0.21, p = 0.021), interest rates ( = -0.34, p = 0.002), and unemployment ( = -0.42, p = 0.001) have significant negative predictive power (R 2 = 0.68). The HDLM achieves a better predictive performance with RMSE = 1.89, MAPE = 4.9% and DA = 83.6% > 23% better than the baseline LSTM and CNN- LSTM settings. The model minimizes prediction error by 36.1% in cases of strong market shocks, which occur under simulated stress conditions, confirming an increased financial strength. Taken together, the results of the studies support the idea that the incorporation of macroeconomic intelligence into hybrid neural systems significantly enhances forecast reliability and stability of the system. The current research, therefore, adds a strong, interpretive, and policy-related framework to the research fields of predictive finance and resilience engineering systems to predict market upheavals.
Suggested Citation
Shaid Hasan & Ismoth Zerine & Md Mainul Islam & Adib Hossain & Khandaker Ataur Rahman & Zulkernain Doha, 2023.
"Predictive Modeling of U.S. Stock Market Trends Using Hybrid Deep Learning and Economic Indicators to Strengthen National Financial Resilience,"
Journal of Economics, Finance and Accounting Studies, Al-Kindi Center for Research and Development, vol. 5(3), pages 223-235, April.
Handle:
RePEc:alk:jefas0:v:5:y:2023:i:3:id:11312
DOI: 10.32996/jefas.2023.5.3.18
Download full text from publisher
Corrections
All material on this site has been provided by the respective publishers and authors. You can help correct errors and omissions. When requesting a correction, please mention this item's handle: RePEc:alk:jefas0:v:5:y:2023:i:3:id:11312. See general information about how to correct material in RePEc.
If you have authored this item and are not yet registered with RePEc, we encourage you to do it here. This allows to link your profile to this item. It also allows you to accept potential citations to this item that we are uncertain about.
We have no bibliographic references for this item. You can help adding them by using this form .
If you know of missing items citing this one, you can help us creating those links by adding the relevant references in the same way as above, for each refering item. If you are a registered author of this item, you may also want to check the "citations" tab in your RePEc Author Service profile, as there may be some citations waiting for confirmation.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: Al-Kindi Center for Research and Development (email available below). General contact details of provider: https://al-kindipublisher.com/index.php/jefas .
Please note that corrections may take a couple of weeks to filter through
the various RePEc services.