INTEGRATION OF PREDICTIVE ARTIFICIAL INTELLIGENCE METHODS INTO INVESTMENT PORTFOLIO DECISION-MAKING

A comprehensive study of artificial intelligence methods suitable for forecasting financial indicators and rational formation of an enterprise's investment portfolio has been conducted, with the systematization of models capable of reflecting nonlinear dependencies, adapting to market volatility, and providing a more accurate assessment of expected profitability and risk. The feasibility of combining forecasting results obtained using neural network algorithms with classical economic and mathematical optimization is substantiated, which allows to increase the reliability of portfolio parameter estimation and ensure a better balance between profitability and risk compared to traditional approaches. Based on the financial data sets, an analysis of the quality of forecasting was carried out, a portfolio was formed considering the established restrictions and the results were evaluated using indicators of stability, risk and generalized efficiency. It has been proven that the use of artificial intelligence methods contributes to increasing the stability of the investment portfolio to market fluctuations, reducing losses during unfavorable phases of market dynamics, and improving overall performance, which indicates the practical value of the author's approach for the corporate financial management system.