EXPLAINABLE AI-BASED DYNAMIC PORTFOLIO OPTIMIZATION OF ISLAMIC EQUITIES, GREEN AND CONVENTIONAL CRYPTOCURRENCIES, GOLD, AND CRUDE OIL: EVIDENCE FROM PAKISTAN AND MALAYSIA

Abstract

The increasing integration of digital assets, sustainable investments, and Islamic financial markets has created new opportunities and challenges for portfolio management. While cryptocurrencies offer high return potential, their significant volatility necessitates advanced risk management techniques. At the same time, Islamic equities and green cryptocurrencies have emerged as attractive investment alternatives for investors seeking ethical and sustainable portfolios. Traditional commodities, such as gold and crude oil, continue to serve as important hedging and diversification instruments during periods of market uncertainty. Despite these developments, limited research has integrated these asset classes within an Explainable Artificial Intelligence (XAI)-based dynamic portfolio optimization framework, particularly in emerging Islamic financial markets. This study addresses this gap by developing an XAI-driven portfolio optimization model that combines Islamic equities, green and conventional cryptocurrencies, gold, and crude oil for Pakistan and Malaysia. Daily price data are analyzed using the Conditional Value-at-Risk (CVaR) optimization approach to minimize downside risk while maximizing risk-adjusted returns. Explainable AI enhances model transparency by providing interpretable insights into portfolio allocation decisions and asset contributions. The findings demonstrate that integrating diverse asset classes improves portfolio diversification and resilience under volatile market conditions. Furthermore, differences in optimal asset allocations between Pakistan and Malaysia highlight the importance of country-specific investment strategies. This study contributes to the literature on intelligent portfolio management, sustainable finance, and Islamic investment by providing a transparent and adaptive framework for informed investment decision-making.