Aligning Language Models with Investor and Market Behavior for Financial Recommendations

2025-10-13

AI Summary

Existing financial recommendation systems often neglect behavioral biases and regulatory constraints, resulting in recommendations misaligned with user preferences, poor interpretability, and low adoption rates. To address these issues, we propose FLARKO—a novel framework that pioneers the integration of Kahneman–Tversky Optimization (KTO) into large language model (LLM) fine-tuning, while incorporating structured knowledge graphs (KGs) within a federated learning paradigm to enhance LLMs’ reasoning over behavioral finance data. FLARKO achieves behavioral alignment, robust investment performance, high interpretability, and low computational overhead under both centralized and federated settings. Extensive experiments on the FAR-Trans dataset demonstrate that FLARKO significantly outperforms state-of-the-art methods, striking a superior trade-off between behavioral consistency and investment efficacy.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Game Theory and Economic Paradigms: Behavioral Game Theory

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Most financial recommendation systems often fail to account for key behavioral and regulatory factors, leading to advice that is misaligned with user preferences, difficult to interpret, or unlikely to be followed. We present FLARKO (Financial Language-model for Asset Recommendation with Knowledge-graph Optimization), a novel framework that integrates Large Language Models (LLMs), Knowledge Graphs (KGs), and Kahneman-Tversky Optimization (KTO) to generate asset recommendations that are both profitable and behaviorally aligned. FLARKO encodes users' transaction histories and asset trends as structured KGs, providing interpretable and controllable context for the LLM. To demonstrate the adaptability of our approach, we develop and evaluate both a centralized architecture (CenFLARKO) and a federated variant (FedFLARKO). To our knowledge, this is the first demonstration of combining KTO for fine-tuning of LLMs for financial asset recommendation. We also present the first use of structured KGs to ground LLM reasoning over behavioral financial data in a federated learning (FL) setting. Evaluated on the FAR-Trans dataset, FLARKO consistently outperforms state-of-the-art recommendation baselines on behavioral alignment and joint profitability, while remaining interpretable and resource-efficient.

Problem

• Financial systems ignore behavioral factors causing misaligned advice
• Generating profitable recommendations aligned with investor preferences
• Providing interpretable asset suggestions using behavioral financial data

Innovation

• Integrates LLMs with knowledge graphs for recommendations
• Uses Kahneman-Tversky Optimization for behavioral alignment
• Implements both centralized and federated learning architectures