用大模型预测金融风险，效果优于传统方法
Time-Series Foundation AI Model for Value-at-Risk Forecasting
- 用预训练时序大模型TimesFM，通过微调提升风险预测精度
- 在超过8.5年测试数据上，模型在0.01至0.1分位数预测中表现最优
- 适合金融风控、量化交易等需要高精度尾部风险预测的场景
本研究首次评估时序基础大模型在价值风险（VaR）预测中的表现，即对收益分布左尾分位数进行预测。基础模型在多样化数据上预训练，可在零样本设置下使用，或通过微调进一步优化。我们对比了谷歌的TimesFM模型与传统参数和非参数模型（如GARCH和广义自回归得分模型GAS），使用标普100指数及其成分股19年的日度收益数据。基于超过8.5年的样本外数据回测显示，微调后的基础模型在实际超出预期比率上持续优于传统方法；在分位数评分损失函数下，其表现可与最佳计量经济模型GAS相当。总体而言，该模型在0.01、0.025、0.05和0.1分位数预测中排名最佳或位居前列。微调显著提升准确率，表明零样本应用并非最优方案。
原文摘要 · Abstract (English)
This study is the first to analyze the performance of a time-series foundation AI model for Value-at-Risk (VaR), which essentially forecasts the left-tail quantiles of returns. Foundation models, pre-trained on diverse datasets, can be applied in a zero-shot setting with minimal data or further improved through finetuning. We compare Google's TimesFM model to conventional parametric and non-parametric models, including GARCH and Generalized Autoregressive Score (GAS), using 19 years of daily returns from the SP 100 index and its constituents. Backtesting with over 8.5 years of out-of-sample data shows that the fine-tuned foundation model consistently outperforms traditional methods in actual-over-expected ratios. For the quantile score loss function, it performs comparably to the best econometric model, GAS. Overall, the foundation model ranks as the best or among the top performers across the 0.01, 0.025, 0.05, and 0.1 quantile forecasting. Fine-tuning significantly improves accuracy, showing that zero-shot use is not optimal for VaR.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。