Stock Price Prediction Model Using LSTM: A Comparative Study

Due to its potential use in financial markets, stock price prediction is a difficult issue that has drawn considerable interest from academics and business executives. Recurrent neural networks (RNNs) with Long Short-Term Memory (LSTM) have demonstrated promising results in time series prediction applications, such as stock price prediction. In this paper, we examine both univariate and multivariate LSTM-based stock price prediction models. We compare the forecasting accuracy of these models using a variety of assessment criteria in order to measure their performance on a real-world stock price dataset. Additionally, we look into how different hyperparameters, such as batch size, learning rate, and the quantity of hidden layers, affect how well LSTM models perform. Finally, we talk about the interpretability, limitations, and prospective uses of LSTM-based stock price prediction models in practical financial decision-making scenarios.