Search papers, labs, and topics across Lattice.
This paper introduces Expert-Following Strategies, a novel framework for financial asset recommendation that leverages historical transaction data to align investment returns with user preferences. By identifying top-performing investors and recommending assets based on ROI-weighted purchase frequency, the approach addresses the critical trade-off between profitability and relevance inherent in existing methods. Experiments demonstrate that this strategy significantly outperforms market-average baselines in both return on investment and normalized discounted cumulative gain across various thresholds.
Expert-Following Strategies simultaneously boost investment returns and relevance, breaking the traditional trade-off in financial recommendations.
Financial institutions hold rich transaction histories, yet delivering recommendations that simultaneously maximize investment returns and ensure preference alignment remains a significant challenge. Existing approaches, namely return-based and preference-based strategies, each optimize a single objective, resulting in a fundamental trade-off between profitability (ROI) and relevance (nDCG). In this paper, we propose the Expert-Following Strategies: a framework that identifies top-performing investors based on their historical ROI and recommends the assets they purchased, scored by ROI-weighted purchase frequency. Our experiments using real-world transaction histories show that our strategy achieves statistically significant improvement over the market-average baseline in both ROI and nDCG simultaneously across all four thresholds.
