Technology Category
Machine Learning: Large Multimodal Models (LMMs)Multiagent Systems: Mechanism DesignCognitive Modeling & Cognitive Systems: Social Cognition And Interaction
PulseReddit: A Novel Reddit Dataset for Benchmarking MAS in High-Frequency Cryptocurrency Trading
🤖 AI Summary
This study investigates whether social media sentiment can enhance high-frequency cryptocurrency trading performance. To this end, we construct the first large-scale, millisecond-aligned dataset integrating Reddit discussion threads with on-chain market data—enabling precise temporal synchronization between textual and market time series. We propose LLM-MAS, a large language model–based multi-agent system that unifies real-time sentiment analysis with high-frequency backtesting, facilitating social-signal–driven trading decisions. Empirical results show that integrating PulseReddit improves the Sharpe ratio by 23.7% and consistently generates positive alpha across multiple market regimes. We further quantify the operational trade-offs between latency and accuracy for models including GPT-4 and Llama-3. Our core contributions are: (1) the first publicly available, temporally aligned Reddit–on-chain dataset; (2) a scalable, modular LLM-MAS architecture; and (3) rigorous empirical validation of social signals’ statistically significant contribution to high-frequency trading returns.
Technology Category
Application Category
📝 Abstract
High-Frequency Trading (HFT) is pivotal in cryptocurrency markets, demanding rapid decision-making. Social media platforms like Reddit offer valuable, yet underexplored, information for such high-frequency, short-term trading. This paper introduces extbf{PulseReddit}, a novel dataset that is the first to align large-scale Reddit discussion data with high-frequency cryptocurrency market statistics for short-term trading analysis. We conduct an extensive empirical study using Large Language Model (LLM)-based Multi-Agent Systems (MAS) to investigate the impact of social sentiment from PulseReddit on trading performance. Our experiments conclude that MAS augmented with PulseReddit data achieve superior trading outcomes compared to traditional baselines, particularly in bull markets, and demonstrate robust adaptability across different market regimes. Furthermore, our research provides conclusive insights into the performance-efficiency trade-offs of different LLMs, detailing significant considerations for practical model selection in HFT applications. PulseReddit and our findings establish a foundation for advanced MAS research in HFT, demonstrating the tangible benefits of integrating social media.
Problem
Research questions and friction points this paper is trying to address.
Aligning Reddit data with high-frequency crypto trading metrics
Assessing social sentiment's impact on trading via LLM-based MAS
Evaluating LLM performance-efficiency trade-offs in HFT applications
Innovation
Methods, ideas, or system contributions that make the work stand out.
PulseReddit dataset aligns Reddit with crypto stats
LLM-based MAS analyzes social sentiment for trading
MAS with PulseReddit outperforms traditional baselines
🔎 Similar Papers
No similar papers found.
