Application of XGBoost in investment fund analysis - selected case studies

This paper surveys and critically evaluates the literature on the application of XGBoost, one of the machine learning algorithms, in the selection of investment funds and the analysis of the determinants of their performance. The article presents the methodological foundations of the algorithm and the growing importance of its use in financial market analysis. It discusses examples from American, Asian, and European markets, demonstrating that XGBoost offers higher predictive accuracy and interpretability compared to traditional statistical methods. However, the study also highlights challenges, including the need to account for broader data contexts, transaction costs, and longer time horizons to avoid overfitting and ensure model robustness.