Stock Market Analysis and Prediction using Machine Learning

This work presents a web-based stock market analysis and prediction system that integrates machine learning models with financial time series and sentiment data to support short-term trading decisions. Historical price data for US and Indian equities are collected from Yahoo Finance and, when necessary, from Alpha Vantage, and are used to train ARIMA, Long Short-Term Memory (LSTM) networks, and Linear Regression models that forecast the next-day closing price and compute root mean squared error as an accuracy metric. In parallel, the system aggregates recent news headlines via news API and Twitter posts via the Twitter API, applies VADER sentiment analysis, and derives an overall polarity score that reflects the current market mood for the selected stock. These predictive and sentiment components are integrated into a Flask application with responsive HTML dashboards that visualize recent price trends, model predictions, error metrics, tweet polarity, and seven-day forecasted prices, and generate an interpretable BUY or SELL recommendation for end-users. The proposed framework demonstrates how combining statistical, deep learning, and sentiment-driven signals in a single platform can enhance the transparency and usability of data-driven stock decision support for retail investors.