Predicting Firms' Performance in GCC Countries Using Data Mining Techniques
| Author | Alanzi, Jawaher Khaled Aref | |
|---|---|---|
| metadata.dc.contributor.department | "University of Bahrain, College of Business Administration, Department of Management and Marketing | |
| metadata.dc.date.accessioned | 2026-09-20T15:54:10Z | |
| Publication date | 2025 | |
| Degree level | Doctoral | |
| Abstract | Abstract : Purpose: This research aims to examine the accuracy of Data Mining (DM) techniques in predicting the financial performance of publicly listed companies in the Gulf Cooperation Council (GCC) countries (namely, the Kingdom of Bahrain, Kuwait, Oman, Qatar, the Kingdom of Saudi Arabia, and the United Arab Emirates (UAE)). Design/Methodology/Approach: The study analyzed previous literature on predicting firms' performance using DM techniques, combining Systematic Literature Review (SLR) and bibliometric analysis to identify literature gaps. A conceptual model was developed to examine the relationship between independent variables: firm-specific (liquidity, leverage, solvency, asset utilization, and market value), macroeconomic (GDP and inflation), and control variables (firm size, age, industry type, and firm's country), and the dependent variable measured by Return On Investment (ROI). The dataset was sourced from the London Stock Exchange Group (LSEG) and the World Development Indicators (WDI) and covered 816 publicly listed companies across diverse industries of the GCC countries from 2003 to 2022. The study evaluated four prediction models using seven data manipulation techniques implemented in Python. These included single data manipulation techniques: Linear Regression (LR), Artificial Neural Networks (ANN), Regression Trees (RT), and k Nearest Neighbors (kNN), along with ensemble models such as Random Forest (RF), eXtreme Gradient Boosting (XGBoost), and Categorical Boosting (CatBoost). Findings: CatBoost outperforms other models, achieving the highest prediction accuracy, with R-squared (0.542) for the all-GCC dataset model. On the other hand, Bahrain's country-specific model demonstrated the best prediction accuracy (R-squared = 0.781, lowest error rates). In contrast, Saudi Arabia's model showed the weakest prediction accuracy (R-squared = 0.328, highest error rates). In terms of variable importance, asset utilization and solvency demonstrate a positive influence on profitability across most GCC countries. These results indicate that firm-specific factors have the most significant influence on profitability. Macroeconomic variables, such as GDP and inflation, have a minimal influence on profitability, whereas control variables (age, industry type, country, and size) exert a moderate effect, with a negative impact of age on profitability. Originality/Value: From a theoretical perspective, this study contributes novelty to the literature by exploring the capabilities of DM techniques in predicting ROI as a financial measure for profitability. This study examines the firm-specific factors that influence publicly listed companies' performance in the GCC region and assesses the macroeconomic variables that affect their performance across diverse industries. The study emphasizes the importance of implementing country-specific DM models rather than using an all-GCC dataset model to account for economic variations across countries. Furthermore, the findings of this research hold substantial empirical significance, offering valuable insights for informed decision-making among various stakeholders. Policymakers and regulators in the GCC, such as the Capital Market Authorities (CMA) and bankruptcy commissions, may utilize this information to take early interventions, preventing business failures and enabling proactive measures for firms identified as being at risk (e.g., low profitability prediction). Additionally, managers can leverage insights based on the importance of independent variables to optimize internal processes and resource allocation, ultimately enhancing financial performance. Furthermore, shareholders can make more informed strategic decisions regarding stock transactions based on the firm's predicted profitability outcomes | |
| metadata.dc.identifier.other | Drupal node 4659 | |
| metadata.dc.identifier.other | https://digitalrepository.uob.edu.bh/en/dar/predicting-firms-performance-gcc-countries-using-data-mining-techniques | |
| metadata.dc.identifier.uri | http://165.227.56.122/handle/123456789/7551 | |
| Language | en | |
| Link | https://digitalrepository.uob.edu.bh/en/_flysystem/fedora/2026-02/002423_predicting_firms_performance_in_gcc_countries_using_data_mining_techniques.pdf | |
| Title | Predicting Firms' Performance in GCC Countries Using Data Mining Techniques | |
| Type | Doctoral Dissertation | |
| College | College of Business Administration | |
| Department | Management and Marketing |
Files
Original bundle
1 - 1 of 1
Loading...
- Name:
- node-4659-1-002423_predicting_firms_performance_in_gcc_countries_using_data_mining_techniques.pdf
- Size:
- 10.1 MB
- Format:
- Adobe Portable Document Format