Stock Price Prediction Using Machine Learning

Abstract

Abstract—As the need for precise and fast stock price projections is growing as financial markets continue to become more complicated. The use of Long Short-Term Memory (LSTM) neural networks for stock price forecasting is investigated in this paper. LSTMs are especially important for time series prediction tasks because they are good at capturing temporal dependencies in sequential data. In this study, LSTM models are trained and evaluated using historical stock price data while taking into account a variety of input features and hyperparameters. By contrasting the performance of LSTM models with conventional machine learning techniques, the advantages and disadvantages of each methodology are brought to light. The findings showed that LSTM models offer greater predictive capabilities, highlighting its potential in improving investment decisions and furthermore, the analysis of IBM Company Stock Price Values and model performance configurations add to the expanding body of knowledge about using deep learning approaches for financial forecasting. Keywords—Machine Learning, Neural Networks, RNN, Linear Regression, LSTM, Supervised Learning.