Stock Market Price Forecasting Using Time-Series ML

Stock market prediction is a challenging task influenced by economic conditions, investor behavior, and global events. Accurate short-term forecasting can assist investors in decision-making and risk management. This study presents a machine learning–based framework for next-day stock price prediction using historical time-series data. Daily trading data from five major Indian companies—TCS, Reliance Industries, HDFC Bank, Infosys, and ICICI Bank—were analyzed to identify price trends and predictive patterns. Multiple regression models, including Linear Regression, Random Forest Regressor, XGBoost, and Support Vector Regressor (SVR), were evaluated using standard performance metrics such as Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and the coefficient of determination (R2R^2R2). Experimental results indicate that Linear Regression consistently achieved the highest predictive accuracy across all datasets, with an R2R^2R2 value close to 0.999, outperforming more complex models. The trained model was deployed using a Flask-based web application that allows users to input daily stock parameters and receive real-time next-day price predictions along with trend indicators. The results demonstrate that well-engineered, lightweight machine learning models can provide effective and computationally efficient solutions for short-term stock price forecasting.