Structured Agentic Workflows for Financial Time-Series Modeling with LLMs and Reflective Feedback

2025-08-19

AI Summary

Financial time series modeling faces a fundamental trade-off among performance, interpretability, and auditability, while existing AutoML frameworks lack adaptability to domain-specific dynamic requirements. To address this, we propose TS-Agent—a modular agent framework that integrates large language models’ reasoning, memory, and code generation capabilities into a closed-loop “plan–execute–feedback” workflow. Leveraging a structured knowledge base and a dedicated planning agent, TS-Agent enables context-aware model selection, code optimization, and hyperparameter tuning, while ensuring traceable decision paths and robust error mitigation. Evaluated on real-world financial forecasting benchmarks and synthetic time series tasks, TS-Agent consistently outperforms state-of-the-art AutoML systems and agent-based baselines across accuracy, robustness, and modeling transparency—demonstrating simultaneous gains in predictive performance and operational accountability.

Technology Category

• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Machine Learning: Time-Series/Data Streams
• Multiagent Systems: Agent Communication

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Time-series data is central to decision-making in financial markets, yet building high-performing, interpretable, and auditable models remains a major challenge. While Automated Machine Learning (AutoML) frameworks streamline model development, they often lack adaptability and responsiveness to domain-specific needs and evolving objectives. Concurrently, Large Language Models (LLMs) have enabled agentic systems capable of reasoning, memory management, and dynamic code generation, offering a path toward more flexible workflow automation. In this paper, we introduce extsf{TS-Agent}, a modular agentic framework designed to automate and enhance time-series modeling workflows for financial applications. The agent formalizes the pipeline as a structured, iterative decision process across three stages: model selection, code refinement, and fine-tuning, guided by contextual reasoning and experimental feedback. Central to our architecture is a planner agent equipped with structured knowledge banks, curated libraries of models and refinement strategies, which guide exploration, while improving interpretability and reducing error propagation. extsf{TS-Agent} supports adaptive learning, robust debugging, and transparent auditing, key requirements for high-stakes environments such as financial services. Empirical evaluations on diverse financial forecasting and synthetic data generation tasks demonstrate that extsf{TS-Agent} consistently outperforms state-of-the-art AutoML and agentic baselines, achieving superior accuracy, robustness, and decision traceability.

Problem

• Automating financial time-series modeling with interpretable agentic workflows
• Addressing AutoML limitations in adaptability and domain-specific responsiveness
• Enhancing model accuracy and auditability through structured iterative refinement

Innovation

• Modular agentic framework for financial time-series modeling
• Structured iterative decision process with three stages
• Planner agent with knowledge banks for exploration guidance