Predicting stock prices in the Pakistan market using machine learning and technical indicators

Abstract

This study uses advanced machine learning models to predict stock prices in the Pakistani stock market using 27 technical indicators. It evaluates the predictive capabilities of four techniques, SVM, LSTM, and Random Forest for binary classification of stock price movements. ANN and SVM show the highest accuracy at 85%, followed by Random Forest at 84% and LSTM at 78%. Key indicators such as %R, Momentum, and Disparity 5 are critical across all models. The research provides valuable insights for investors and analysts to improve decision-making. It underscores the importance of technical indicators and establishes a data-driven approach to navigating the complexities of the Pakistani stock market. The study further emphasizes the importance of technical indicators and suggests exploring hybrid models that incorporate real-time data, sentiment analysis, and external factors for better stock price prediction.

Keywords

• Machine learning
• Stock prices Forecasting
• Artificial Neural Networks (ANN)
• Support Vector Machines (SVM)
• Long Short-Term Memory (LSTM)
• Random Forest