arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2403.08754math.PR
收藏

带粘性阈值的扩散过程、局部时逼近与参数估计

Sticky-threshold diffusions, local time approximation and parameter estimation

Universidad de Chile(智利大学);Université de Lorraine(洛林大学)

Alexis Anagnostakis, Sara Mazzonetto

AI总结:

该研究针对带奇异阈值或边界的一维扩散过程,构造收敛到局部时的高频路径泛函,结合占据时间逼近得到扩散系数存在跳跃不连续性时仍有效的粘性与偏度参数一致估计量,解决了多奇异特征共存下的参数估计问题。

摘要:

我们研究一类一维扩散过程的高频路径泛函,这类扩散过程带有奇异阈值或边界,允许存在偏度、扩散系数的不连续性以及粘性或粘性反射现象。这些泛函最初是为非奇异扩散过程的局部时逼近而开发的,由一个测试函数和一个发散的归一化序列构造而成。我们证明,无论奇异阈值特征的性质如何,这类泛函都会收敛到局部时,从而推广了近期针对特定情形的若干结果。值得注意的是,我们的框架允许任意满足$o(n)$的发散归一化序列(其中$n$为观测频率),以及同时出现多种奇异行为的阈值。结合局部时逼近与占据时间逼近,我们构造了粘性参数和偏度参数的一致估计量,该估计量在扩散系数存在跳跃不连续性时仍有效;这解决了所有这些奇异特征同时存在时的估计问题。

英文摘要:

We study a class of high-frequency path functionals for one-dimensional diffusions with singular thresholds or boundaries, allowing for skewness, discontinuities in the diffusion coefficient, and stickiness or sticky reflection. These functionals, originally developed for local-time approximation in non-singular diffusions, are constructed from a test function and a diverging normalizing sequence. We establish convergence to local time, regardless of the nature of the singular threshold features, thereby extending several recent results on specific cases. Notably, our framework allows for any normalizing diverging sequence that is $o(n)$, where $n$ is the observation frequency, and thresholds at which several singular behaviors occur simultaneously. Combining the local-time approximation with occupation-time approximations, we construct consistent estimators of the stickiness and skewness parameters that remain valid in the presence of jump-discontinuities in the diffusion coefficient; this solves the estimation problem under the simultaneous presence of all these singular features.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑