Stock Price Prediction Web Application Using LSTM and Streamlit

This paper presents a comprehensive web-based application for stock price prediction using Long Short-Term Memory (LSTM) neural networks. The system is designed to forecast future stock prices based on historical financial data and provides users with interactive visualizations and predictive analytics. The frontend is developed with Streamlit for real-time interaction, while the backend leverages TensorFlow and financial APIs to fetch, preprocess, and model stock data. The model demonstrates robust performance in short-term forecasting, with Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE) used as key evaluation metrics. This application offers a user-friendly interface and supports features such as customizable prediction date ranges, CSV data uploads, model retraining, and dynamic charting. The goal is to democratize access to predictive financial modeling and provide an accessible tool for investors, students, and researchers.