跳至正文
返回文库全部文档

使用分位数模型概率预测加密货币波动率

文章 arXiv papers · 作者: Grzegorz Dudek et al.
阅读原文

总结

本研究构建比特币已实现方差的概率预测模型，不局限于点预测，还估计条件分位数，以表示可能波动结果的不确定性。研究结合 HAR、GARCH 和 ARFIMA 等统计模型与机器学习方法的预测，再通过概率堆叠框架整合基础模型的预测结果。

报告的比较结果显示，残差模拟分位数估计法表现持续良好，尤其是与基于对数转换已实现波动率训练的线性基础模型结合时。本文据此指出，即使面对更复杂的替代方法，较简单的线性预测输入也能支持有效的概率估计。分析仅针对比特币，所提供的描述没有给出样本时期、评估指标或模型设置详情。因此，这些发现为风险感知型波动率预测提供了参考，但不能证明其他加密货币或市场条件下也会得到相同的模型排序。

核心观点

  • 概率预测估计条件已实现方差的多个可能结果，而非只给出单一数值。
  • 该框架结合统计和机器学习基础模型的预测。
  • 据报告，残差模拟法表现持续良好，尤其是结合使用对数波动率数据的线性模型时。
  • 实证发现聚焦于比特币，不能证明其在其他加密货币上的表现。

标签

加密资产波动率机器学习统计学风险管理
全文
# Probabilistic Forecasting Cryptocurrencies Volatility: From Point to Quantile Forecasts


# Probabilistic Forecasting Cryptocurrencies Volatility: From Point to Quantile Forecasts









Cryptocurrency markets are characterized by extreme volatility, making accurate forecasts essential for effective risk management and informed trading strategies. Traditional deterministic (point) forecasting methods are inadequate for capturing the full spectrum of potential volatility outcomes, underscoring the importance of probabilistic approaches. To address this limitation, this paper introduces probabilistic forecasting methods that leverage point forecasts from a wide range of base models, including statistical (HAR, GARCH, ARFIMA) and machine learning (e.g. LASSO, SVR, MLP, Random Forest, LSTM) algorithms, to estimate conditional quantiles of cryptocurrency realized variance. To the best of our knowledge, this is the first study in the literature to propose and systematically evaluate probabilistic forecasts of variance in cryptocurrency markets based on predictions derived from multiple base models. Our empirical results for Bitcoin demonstrate that the Quantile Estimation through Residual Simulation (QRS) method, particularly when applied to linear base models operating on log-transformed realized volatility data, consistently outperforms more sophisticated alternatives. Additionally, we highlight the robustness of the probabilistic stacking framework, providing comprehensive insights into uncertainty and risk inherent in cryptocurrency volatility forecasting. This research fills a significant gap in the literature, contributing practical probabilistic forecasting methodologies tailored specifically to cryptocurrency markets.

在遵守原作品许可的前提下，附作者信息全文展示。 许可协议: abstract CC0

此摘要由 Stratmill 研究智能体根据原文撰写，并非原文副本。