Abstract This study presents a narrative comparative analysis of Machine Learning (ML) techniques and traditional financial models for portfolio optimization in emerging markets. Portfolio optimization has long relied on conventional approaches such as the Mean-Variance Model, Capital Asset Pricing Model (CAPM), and Modern Portfolio Theory (MPT), which are based on assumptions regarding market efficiency, risk-return relationships, and historical data patterns. However, the increasing availability of large financial datasets and advancements in computational technologies have led to the adoption of Machine Learning methods capable of identifying complex, nonlinear relationships within financial markets. The study reviews and compares the theoretical foundations, methodologies, advantages, and limitations of both approaches in the context of emerging markets, which are often characterized by higher volatility, market inefficiencies, information asymmetry, and rapidly changing economic conditions. Machine Learning models, including Artificial Neural Networks, Random Forests, Support Vector Machines, and Deep Learning techniques, demonstrate strong predictive capabilities and adaptability to dynamic market environments. In contrast, traditional financial models offer greater interpretability, simplicity, and established theoretical frameworks but may struggle to capture complex market behaviors. The findings suggest that while Machine Learning approaches frequently outperform traditional models in forecasting and risk management, their effectiveness depends on data quality, computational resources, and model transparency. The study concludes that a hybrid approach integrating Machine Learning techniques with traditional financial theories may provide a more robust framework for portfolio optimization in emerging markets. Such integration can enhance investment decision-making, improve risk-adjusted returns, and support more efficient portfolio management practices.
Maitrayee Ashok Divekar (Thu,) studied this question.