GPD Threshold Estimation Using Measure of Surprise

Abraham Manurung, Aji Hamim Wigena, Anik Djuraidah

Abstract


Threshold is used to estimate parameters of Generalized Pareto distribution to estimate return value. This return value shows the extreme value in the period of time. Threshold can be estimated using Mean Residual Life Plot, Threshold Stability Plot, or the upper 10% rule but this estimation is usually subjective. An alternative method is measure of surprise based on Bayes method and Monte Carlo Markov Chain technique. This paper aims to estimate a GPD threshold based on simulation data and to apply measure of surprise method to rainfall data in the period of 1981-2012 in Bogor, Indonesia.  The simulation result showed that the predicted threshold is exactly the same as the true threshold. The result of application to rainfall data showed the threshold was approximately 210 mm.


Keywords


Bayes; GPD; measure of surprise; posterior predictive distribution; rainfall; threshold estimation.

Full Text:

PDF

References


S. Coles. An Introduction to Statistical Modeling of Extreme Values. London, UK: Springer, 2001.

I. Vicari. “Measure of surprise and threshold selection in extreme value statistics”. Thesis. École Polytechnique Fédérale de Lausanne, 2010.

J. Pickands. “Statistical inference using extreme order statistics”. Annals of Statistics, vol. 3, pp. 119-131, 1975.

A.C. Davison, R.L. Smith. “Models for exceedances over high thresholds”. Journal of the Royal Statistical Society, Series B, vol. 52, pp 393-442, 1990.

C.N. Behrens, F.L. Hedibert, and D. Gamerman. “Bayesian analysis of extreme events with threshold estimation”. Statistical Modelling, vol. 4, pp. 227-244, 2004.

C.J. Scarrott, A. MacDonald. “A review of extreme value threshold estimation and uncertainty quantification”. Statistical Journal, vol. 10, pp. 33-60, 2012.

W.H. DuMouchel. “Estimating the stable index α in order to measure tail thickness: A Critique”. The Anns of Statistics, vol.11, pp.1019-1031, 1983

W. Weaver. “Probability, rarity, interest and surprise”. Scientific Monthly. vol. 67, pp. 390-392, 1948.

G.E.P Box. “Sampling and bayes inference in scientific modeling and robustness”. Journal of the Royal Statistical Society, Series A, vol. 143, pp. 383-430, 1980.

M. J. Bayarri, J.O.Berger. “Measure of Surprise in Bayesian Analysis”. ISDS Discussion Paper, 1997, pp. 97-154.

X. L. Meng. “Posterior predictive p-values”. The Annals of Statistics, vol. 22, pp. 1142-1160, 1994.

J. Lee, Y. Fan, S.A. Sisson. “Bayesian threshold selection for extremal models using measures of surprise”. Computational Statistics and Data Analysis, vol. 85, pp. 84-99, 2015.

I. Guttman. “The use of the concept of a future observation in goodness-of- fit problems”. Journal of the Royal Statistical Society, Series B, vol. 29, pp. 83-100, 1967.

M. Castellanos S. Cabras, J. “A default Bayesian procedure for the generalized Pareto distribution”. Journal of Statistical Planning and Inference, vol. 137, pp. 473-483, 2007.

D.B. Stephenson, K.R. Kumar, F.J. Doblas-Reyes, J.F. Royer, E. Chauvin, S. Pezzulli. “Extreme Daily Rainfall Events and Their Impact on Ensemble Forecasts of the Indian Monsoon”. Monthy Weather Review, vol. 127, pp. 1955-1966, 1999.


Refbacks

  • There are currently no refbacks.


 

 
  
 

 

  


About IJSBAR | Privacy PolicyTerms & Conditions | Contact Us | DisclaimerFAQs 

IJSBAR is published by (GSSRR).