Kalman-Based Forecasting of Stock Market Trend

Abstract

This paper proposes a Kalman filter (KF)-based algorithm to predict the stock market price. The financial data of certain companies listed in the Bombay Stock Exchange (BSE) India for the year 2012 from the official website of BSE were used. The prices of the stock for the past days along with certain technical indices such as the moving average convergence/divergence (MACD), relative strength index (RSI), and stock momentum (SM) were used for the prediction of the price for the next day. The authors first propose a new model wherein the highest stock price is obtained as the linear combination of the MACD, RDI, and SM. Following this, the KF recursively updates the parameters upon the completion of the prediction for the corresponding day. The performance of the algorithm is validated with real-data of stock market for five Indian companies i.e., NTPC, IOCL, Coal India limited, GAIL, and BHEL. © 2025 IEEE.