A hybrid artificial intelligence - based approach to stock price prediction and portfolio optimization: A case study on Borsa İstanbul

In today's world, globalization and digital transformation processes are increasing volatility in financial markets, necessitating data-driven approaches that go beyond traditional financial theories to enable investors to make informed decisions. In this context, this study analyzes historical data from stocks listed on the Borsa Istanbul BIST 30 index to generate price forecasts and determine an optimal portfolio composition. In the first phase of the study, four different models—Random Forest, Extreme Gradient Boosting (XGBoost), Support Vector Regression (SVR), and Long Short-Term Memory (LSTM)—were employed to predict stock prices, and their performances were evaluated using error metrics. The analysis revealed that the Random Forest algorithm outperformed the other techniques, achieving higher accuracy rates. Notably, stocks such as AEFES.IS, ASELS.IS, and THYAO.IS demonstrated particularly high prediction performance in the stock-specific analyses. In the first phase of the study, four different models—Random Forest, Extreme Gradient Boosting (XGBoost), Support Vector Regression (SVR), and Long Short-Term Memory (LSTM)—were employed to predict stock prices, and their performances were evaluated using error metrics. The analysis revealed that the Random Forest algorithm outperformed the other techniques, achieving higher accuracy rates. Notably, stocks such as AEFES.IS, ASELS.IS, and THYAO.IS demonstrated particularly high prediction performance in the stock-specific analyses. The findings indicate that artificial intelligence and optimization algorithms significantly enhance decision-making processes in financial markets by minimizing investment risks and maximizing return potential. In addition to enabling investors to make more informed and data-driven decisions, this study makes a valuable contribution to the financial literature. The increasing adoption of AI-based financial analysis is expected to improve market predictability and allow investors to adapt more flexibly and consciously to volatile market conditions. This research serves as an important resource for academics, analysts, and investors who seek to better understand the role of artificial intelligence in financial markets and improve the accuracy of investment decisions. Looking ahead, studies that explore how deep learning and hybrid optimization models can be more effectively applied in financial markets are anticipated to further enhance financial decision-making processes. In this direction, the integration of artificial intelligence and data analytics techniques into the financial sector will continue to be a fundamental component of modern investment strategies.