INFORMATION-THEORETIC ANALYSIS OF STOCHASTIC VOLATILITY MODELS

被引：0

作者：

Pfante, Oliver ^{[1
]}

Bertschinger, Nils ^{[1
]}

机构：

[1] Frankfurt Inst Adv Studies, Syst Risk Grp, D-60438 Frankfurt, Hesse, Germany

来源：

ADVANCES IN COMPLEX SYSTEMS | 2019年 / 22卷 / 01期

基金：

欧洲研究理事会;

关键词：

Information theory; stochastic volatility; Bayesian analysis; MAXIMUM-LIKELIHOOD-ESTIMATION; OPTIONS;

D O I：

10.1142/S021952591850025X

中图分类号：

O1 [数学];

学科分类号：

0701 ; 070101 ;

摘要：

Stochastic volatility models describe asset prices S-t as driven by an unobserved process capturing the random dynamics of volatility sigma(t). We quantify how much information about sigma(t) can be inferred from asset prices S-t in terms of Shannon's mutual information in a twofold way: theoretically, by means of a thorough study of Heston's model; from a machine learning perspective, by means of investigating a family of exponential Ornstein-Uhlenbeck (OU) processes fitted on S&P 500 data.

引用

页数：21

共 32 条

[1] Maximum likelihood estimation of stochastic volatility models [J].