The article examines the comprehensive methodological support for forecasting changes in the value of the investment model portfolio, considering modern challenges of the global financial environment. It is determined that the increasing role of institutional investors and open funds in the international capital market raises the necessity for the development of accurate, adaptive, and explainable forecasting models. Structural changes in the composition of global investment assets have been analyzed, particularly the dominance of the equity component, which leads to increased volatility and complicates the forecasting of portfolio values. Special attention is given to comparing the efficiency of traditional econometric models (ARIMA, VAR, GARCH, LASSO) and modern machine learning methods (Random Forest, XGBoost, LSTM) in forecasting returns and risks. A substantiation for the use of ensemble approaches to modeling, which combine the advantages of various methods and allow for achieving better stability of results, is presented. An integrated methodological approach has been proposed, which encompasses the preparation and structuring of data, the construction of multiple models, their evaluation based on statistical and economic criteria, software implementation, and the application of results in practical portfolio management. The risks and limitations associated with the impact of changing market regimes (concept drift) and the complexity of interpreting deep learning models are also outlined, requiring regular updates of models and the application of XAI technologies. The scientific novelty of the research lies in the formalization of a methodology that considers not only predictive accuracy but also adaptability to market changes, the possibility of retraining, and tools for practical implementation in the processes of strategic asset management. The obtained results confirm the efficiency of a comprehensive approach to portfolio value forecasting as a tool for enhancing the accuracy of managerial decisions in investment activities.
Kuznetsova et al. (Wed,) studied this question.