We propose an ASHARV-MIDAS model that incorporates the asymmetric and long-memory characteristics of financial asset returns, while integrating current return information into the volatility equation to enhance prediction accuracy. Additionally, we derive the lag order expression and conditional variance of short-term volatility in the novel model to analyze its distinction from the classical GARCH-MIDAS model that does not consider current return information. Empirical and robustness tests demonstrate superior in-sample parameter estimation performance and more precise out-of-sample volatility prediction capabilities of our proposed model.

中文翻译：

---

不对称、长记忆和当前回报信息能否改善原油波动预测？ ——来自ASHARV-MIDAS模型的证据

我们提出了一种ASHARV-MIDAS模型，该模型结合了金融资产收益的不对称和长记忆特性，同时将当前收益信息整合到波动率方程中以提高预测准确性。此外，我们推导了新模型中短期波动性的滞后阶表达式和条件方差，以分析其与不考虑当前收益信息的经典GARCH-MIDAS模型的区别。经验和稳健性测试证明了我们提出的模型具有卓越的样本内参数估计性能和更精确的样本外波动率预测能力。