ARCHIVES

Digital financial platforms have made investment accessible to a wide range of users, but decision-making remains difficult due to changing market conditions and uncertainty in data availability. This paper presents an explainable investment decision support system that generates risk-aware portfolio recommendations. The system is designed to remain functional even when real-time data is unavailable. A modular backend processes market and macroeconomic indicators using a structured feature engineering pipeline based on trend, volatility, and momentum. These features are used in a rule-based multi-factor model to estimate risk and determine portfolio allocation. The system also includes a fallback mechanism that ensures continuous operation by using proxy data when live data feeds fail. The output provides key financial insights such as expected return, risk level, and diversification. In addition, the system generates simple explanations and scenario-based regret analysis to improve user understanding. Experimental evaluation using historical data shows that the system produces stable and consistent results across different conditions. The proposed approach focuses on reliability, transparency, and practical usability, making it suitable for real-world financial decision-making applications.