Expert-Following Strategy Boosts ROI and Relevance in Financial Recommendations

A team of researchers from Hokkaido University has introduced Expert-Following Strategies (EFS), a novel framework for financial asset recommendation that bridges the gap between profit maximization and user preference alignment. Traditional approaches typically optimize either return (ROI) or relevance (nDCG), leading to a fundamental trade-off where high-return recommendations often ignore individual preferences, and preference-based methods sacrifice profitability. The paper, submitted to arXiv, highlights this challenge using real-world transaction histories from financial institutions.

The proposed EFS framework identifies top-performing investors based on their historical ROI, then recommends the assets those experts purchased, weighted by ROI-adjusted purchase frequency. This method statistically significantly outperforms the market-average baseline in both ROI and nDCG simultaneously across all four tested thresholds, as demonstrated in their experiments. The study shows that mimicking expert behavior with ROI weighting can align profitability with user satisfaction, offering a practical solution for financial recommendation systems that need to balance competing objectives.