Abstract

The stock price incorporates variables rate of economic growth, inflation rate, overall economy, trade balance, and monetary system that affect the whole stock market. For investors, the principle of the stock price trend has often been unclear due to numerous significant variables. In developing an investment plan or deciding duration for the purchasing or selling of a stock, the prediction of stock markets provides a crucial function. The stock index's non-linear and dynamic nature estimates the stock market avalue is challenging. Deep learning strategies have emerged as a critical technique in the analysis of dynamic temporal data relations. Several studies of deep learning techniques have been effective in making such a prediction. The Long Short Term Memory (LSTM) has gained popularity for estimating stock market prices. LSTM is a particular form of recurrent neural network (RNN) which implements a gradient descent technique. This paper extensively investigates approaches used for stock market forecasts using LSTM, explains them, and conducts a comparative analysis. The stock market's principal application comprises stock price forecasting, index modeling, risk assessment, and return estimates. We include future directions and summarize the importance of applying LSTM for stock market prediction based on our surveyed papers.

Key words: Long short-term memory (LSTM), recurrent neural network (RNN), nifty 50, root mean square error (RMSE), prediction, stock prices.