Development of an Intelligent System for Stock Market Prediction Using Enhanced Deep Learning Technique With Banking Data

Introduction

Government expenditure is prioritised in order to boost economies and raise living standards. Large firms are established in the modern economy so that people may take advantage of globalisation and its fast-economic developments (Jiang 2021). Private stock exchanges, open stock exchanges, and mixed ownership stock exchanges are all types of the stock market (Li, & Pan, 2022). Shares in private corporations are traded on the private stock exchange, whereas publicly traded stocks are traded on the open stock exchange. Companies with mixed ownership have stock that can only be traded to a limited extent on public exchanges (Mehtab & Sen 2020). British and American stock exchanges, such as the responsible for their development (Hu et al., 2021).

When a company needs money for things like growth or debt repayment, it can “go public,” issuing shares of stock that can be exchanged on the secondary market, also called a stock exchange (Mehtab et al., 2021). The company may avoid the possibility of loss, debt, and interest payments by issuing shares in exchange for financial backing. Second, to generate revenue and profit for the benefit of stockholders (Nabipour et al., 2020). These stockholders, also known as investors, can earn a profit in one of two ways: by receiving dividend payments from the firms in which they invested, or by selling their shares of stock at a price higher than the one at which they were originally acquired. Therefore, persons who interest in trying to anticipate changes in stock prices.

Financial organisations, corporations, and individual investors all confront a difficult dilemma when trying to predict stock prices (Shen & Shafiq 2020). Every economy relies on the stock market, and as the main goals of any investment are to maximise profit while minimising risk, it stands to reason that governments should work to improve their stock markets in order to boost their economies (Ji et al., 2021). Predicting the direction of the stock market is one of the most effective ways to generate a profit because of the potential for rapid returns on stock market investments. The nonlinear nature of stock market prediction makes it more difficult to foresee how a company's shares will perform in a given market (Liu & Long 2020). Therefore, it is necessary for researchers and investors to seek out methods that may produce more reliable outcomes and more revenues.

Stock market prediction refers to any attempt to foretell or anticipate the future value of a stock, market segment, or market as a whole. Many different types of organisations and people (such as traders, market participants, data analysts, computer engineers specialising in ML and AI, etc.) have been fixated on this topic in recent years (Rouf et al., 2021). Due to the fact that the market value of a company's shares is highly dependent on its profits and performance in the marketplace, the value of those shares can fluctuate depending on macroeconomic and microeconomic factors like supply and demand (Mehtab & Sen 2020). An investor may avoid losing money and maximise earnings with the use of such systems and software, which can help them foresee the company's status based on historical and current data, market conditions, etc.

Statistical methods like ARIMA are inferior to traditional models (Jing et al., 2021). However, it has been demonstrated that models such as LSTM are superior to machine learning models such as Regression, and that the deep learning model was detected rather than the machine learning model Support Vector Machine (SVM) (Long et al., 2020). In several cases, deep learning models performed exceptionally well. Due to their capacity to recognise the stock market fluctuations and provide sufficient findings, they shown promise for application in stock market prediction (Ravikumar & Saraf 2020).