J. L. Romero, J. M. Angulo
The compound distribution function of a random field plays a key role in spatial risk assessment. It is directly involved in the specification of the threshold corresponding to an expected relative exceedance area, and conversely. From a methodological point of view, the empirical derivation of this function allows the evaluation of risk measures for diverse exceedance random indicators. In practice, it is often interesting to formulate smooth representations for a given real phenomenon, suitable for analysis at different scales of resolution. This may be motivated by specific objectives of the study, or related to the complexity inherent to a possibly locally non-regular nature of the underlying random field. In this sense, the use of regularising sequences constitutes a useful approach under adequate convergence properties. This work is focussed on the asymptotic behaviour of these transformations in reference to the different elements involved in spatial risk assessment.
Keywords: extremal behaviour, measures of risk, regularizing sequences, spatial and spatiotemporal random fields, threshold exceedances.
Scheduled
Spatial and Spatio-Temporal Statistics
November 7, 2023 4:50 PM
CC3: Room 1