Abstract

Purpose:

This study aims to evaluate the predicting ability of two machine learning models (ANN and KNN) to predict the stock prices of GCC Islamic banks.

Methodology:

The study analyzes the accuracy rates of two machine learning models to predict the stock prices of GCC Islamic banks by using the daily stock prices (open, close, high, and low) for a period of ten years (May 4th, 2012–May 31st, 2022).

Findings:

The predicting ability of the trained ANN and KNN models is evaluated using standard strategic indicators: root mean square error (RMSE), mean bias error (MBE), and accuracy rates. All the evaluated statistics confirmed that the trained KNN model has better predicting ability compared to ANN, with the highest accuracy level of 99.96% and lowest error values.

Significance:

The contributions of this work to management literature include adding to the scarce work on management interventions to increase altruism in the workforce and the factors that play a role in this process.

Originality:

This study is distinctive in that it emphasizes the significance of machine learning models as an alternative to traditional methods for predicting stock prices.

Research Limitations:

The study has conducted a comparative analysis of the accuracy rates of only two machine learning models and determined the superior model among them. Additional machine learning models can be utilized to evaluate and compare their performances in order to determine the optimal model.

Practical Implications:

The findings of this study have certain implications, as the results can be used to gain significant profits in the GCC region, as they help in predicting market fluctuations, discovering hidden relationships in the data sets, and making accurate investment decisions.

Keywords

Artificial intelligence, Artificial Neural Network, K-Nearest Neighbor, Stock Prediction, MAPE, RMSE, MBE