Modeling multivariate cyber risks: deep learning dating extreme value theory

• DOI: 10.1080/02664763.2021.1936468
• Publication Date: 2021-06-04T13:16:18Z
• Authors: Mingyue Zhang Wu; Jinzhu Luo; Xing Fang; Maochao Xu; Peng Zhao
• Keywords: Quantile

Modeling cyber risks has been an important but challenging task in the domain of security, which is mainly caused by high dimensionality and heavy tails risk patterns. Those obstacles have hindered development statistical modeling multivariate risks. In this work, we propose a novel approach for relies on deep learning extreme value theory. The proposed model not only enjoys accurate point predictions via also can provide satisfactory quantile Both simulation empirical studies show that very well prediction performances.