Price Prediction Using LSTM Based Machine Learning Models

Abstract

The estimation of possible fluctuations in stock prices has been the focus of a lot of research work. Price prediction is a technique for predicting a stock's potential future price, and as a result, the price. This study shows how we can use Machine Learning Models based on Long Short-Term Memory (LSTM) to forecast the price of a stock. Stock prices may be anticipated with a high degree of accuracy if correctly modeled, according to certain suggestions. There is also a lot of literature on basic analysis of stock prices, which focuses on detecting and learning from trends in stock price movements. The focus of this research is on stock market forecasting utilizing Long Short-Term Memory (LSTM) models. For the purpose of our study, we have used DSE30's top 10 companies' historical data. We have built two LSTM models to predict and compare the results of the prediction. To train these models, we used training data that consisted of these companies' stock records from January, 2019 till January, 2021. Our target was to find out which version of the LSTM architecture model gives the best prediction among these models.