A data-driven prediction method for multi-period portfolio optimization using the real options approach

Financial portfolio optimization balances risk and returns. Traditional multi-period models ignore financial time series dynamics and volatility by assuming normally distributed returns and static predictions. Many models ignore unequal estimation penalties, making them difficult. Different distribution models and uncertainty management in finance are sought to fill this gap. We test t-distributions and kernel estimators and add probabilistic risk criteria to the multi-period capital portfolio selection algorithm. Real options manage uncertainty in complex environments and provide accurate forecasts with strong decision-making tools despite volatile financial data. Modern theory applied to empirical applications improves dynamic financial system portfolio optimization and adaptive approaches.