Abstract
In the peak over threshold model resulting in the Extreme-value distribution, type I, (EV1) the firste of the distribution function is based on the Poisson number of exceedances, and the seconde arises from the Exponentially distributed magnitudes.
This paper, on the one hand, generalises the Poisson model to the (positive and negative) Binomial distribution, and, on the other hand, the Exponential distribution is generalised to the Generalised Pareto distribution. Lack of fit with respect to the Poisson and Exponential distribution is measured by statistics derived from those which would be locally most powerful if the estimates of the location and scale parameter were equal to the true parameter values. Ways of combining both statistics are discussed.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
Cox, D.R.; Hinkley, D.V. 1974: Theoretical Statistics. London: Chapman and Hall
Van Montfort, M.A.J.; Gomes, M.I. 1985: Statistical choice of extremal models for complete and censored data. J. Hydrol. 77, 77–87
Revfeim, K.J.A. 1989: A framework for interpreting rainfall models. In: Sansom, J. (ed.) Proceedings Fourth International Meeting on Statistical Climatology, pp. 233–237. 27–31 March, Rotorua, New Zealand
Tiago de Oliveira, J. 1981: Statistical choice of univariate extreme value models. In: Taillie, G. et al. (eds.) Statistical distributions in scientific work, Vol. 6, pp. 367–387. Dordrecht: D. Reidel
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
van Montfort, M.A.J., Otten, A. The first and the second e of the extreme value distribution, EV1. Stochastic Hydrol Hydraul 5, 69–76 (1991). https://doi.org/10.1007/BF01544179
Accepted:
Issue Date:
DOI: https://doi.org/10.1007/BF01544179