Stock Market Prediction Using Machine Learning and Deep Learning Techniques

Abstract

Developing reliable models for Stock Market Analysis can aid investors in making more informed decisions, thereby reducing investment risks and enabling traders to se- lect companies with the best dividends. However, analyzing stock market is challenging due to the significant correlations between stock prices, and traditional batch process- ing methods complicate this task further. This article proposes a novel framework for predicting the closing prices of Tesla and Apple using deep learning techniques, specifi- cally LSTM and a hybrid model combining with CNN with LSTM. The predictions were based on three years of collected data, and the models’ performance was evaluated using MSE,RMSE, NRMSE, and Pearson’s Correlation (R). The stacked LSTM model outper- formed the single CNN-LSTM , achieving an R-squared value of 98.17%. Despite this, the CNN-LSTM model demonstrated superior performance in predicting stock market prices.