Empirical Analysis for Stock Price Prediction Using NARX Model with Exogenous Technical Indicators.

Area of Science:

• Computational Finance
• Artificial Intelligence in Finance
• Time Series Forecasting

Background:

• Accurate stock price prediction is crucial for investment success.
• Artificial intelligence models are increasingly explored for financial market forecasting.
• Predicting daily stock prices for specific companies like CIMB presents unique challenges.

Purpose of the Study:

• To investigate the prediction of daily stock prices for Commerce International Merchant Bankers (CIMB).
• To evaluate the effectiveness of technical indicators within a NARX neural network model.
• To identify optimal artificial neural network (ANN) parameters for enhanced prediction accuracy.

Main Methods:

• Utilized a NARX (Nonlinear Autoregressive with Exogenous Inputs) neural network model.
• Employed technical indicators as input variables for the NARX model.
• Conducted comprehensive parameter tuning and optimization for network configurations and input/output parameters.

Main Results:

• The NARX model incorporating technical indicators demonstrated improved one-step-ahead prediction accuracy for CIMB stock.
• Optimization of input data and neural network parameters further enhanced the model's predictive performance.
• Performance was assessed using metrics including Mean Squared Error (MSE), R-squared, and hit rate.

Conclusions:

• Technical indicators combined with a tuned NARX neural network offer a superior approach to predicting CIMB stock prices.
• Optimized AI models can significantly enhance the accuracy of financial market predictions.
• Improved stock price forecasting can potentially lead to increased investor returns.