Stock Market Prediction During Covid Using LSTM

Abstract

Researchers have always faced significant challenges when attempting to predict stock market movements in the field of computation. This difficulty arises due to the influential nature of stock prices, which are impacted by a multitude of factors, including both tangible and intangible elements such as physical and physiological factors, rational and irrational behavior, geopolitical stability, and investor sentiment. Successful investors strive to anticipate future market conditions for profitable investments. In light of these considerations, we propose the utilization of a stacked long-short-term-memory (LSTM) model to forecast the closing index of stock prices during the uncertain period of the pandemic. The model's performance is evaluated using the root mean square error (RMSE) as a performance metric. Our objective is to optimize the model to enhance prediction accuracy and achieve superior stock market forecasting. The dataset employed in this study encompasses stock market data from NIFTY 50 (India), DAX 40 Index (Germany), FTSE Indices (UK), and S&P 500 Indices (USA), spanning across four sectors: Banking, Information Technology, Healthcare, and Retail. The duration of the dataset ranges from January 30, 2020, to March 31, 2022. The primary goal of this research paper is to analyze historical data and extract future patterns and insights.