@unpublished{epos1321,
          author = {Andrzej Kijko},
         address = {Africa},
           title = {Flaw in the EPRI Procedure 
 for the Maximum Earthquake Magnitude Estimation 
 and its Correction},
            type = {Documentation},
       publisher = {None},
     institution = {University of Pretoria},
         journal = {Flaw in the EPRI Procedure   for the Maximum Earthquake Magnitude Estimation   and its Correction},
            year = {2009},
             url = {https://episodesplatform.eu/eprints/1321/},
        abstract = {It has been shown that the EPRI, Bayesian procedure for the estimation of the regional,  maximum  possible  earthquake  magnitude  m\_max ,  (Cornell,  1994;  EPRI  NP-
4726), provides value of m{\^{ }}\_max, which is biased and systematically underestimated. The value of underestimation can be as big as 0.5 unit of magnitude. The bias is caused by the fact that one of the components of the posterior likelihood function is the sample likelihood  function,  for  which,  the  range  of  observations  (earthquake  magnitudes) depends on the unknown parameter m max . Such dependence violates the rule of optimal properties of the maximum likelihood estimators: so the built likelihood function has its maximum  at  the  maximum  observed  earthquake  magnitude  obs\_m\_max,  not  at  required 
maximum possible magnitude m max. A possible correction to the current EPRI procedure 
is also presented.}
}