[R] Extreme value probabilities

Sarah Henderson sarah.b.henderson at gmail.com
Thu Dec 2 20:33:32 CET 2010


Greetings to all --

I have a set of ~1300 radon measurements from homes across British
Columbia in Canada.  We have split these measurements into five groups
according the the tectonic belt in which they fall, with N ranging
from ~160 to ~600 measurements per group.  The values are roughly
log-normally distributed (1) overall and (2) within each group.

For those areas where values are low I would like to be able to
estimate the probability of finding values greater than 200 Bq/m3 (the
new safety standard) even though the maximum measured value is 174
Bq/m3.  I know this is a question of extreme value distributions, and
the reading I have done so far suggests that there are multiple
approaches (maximum likelihood, probability-weighted moments,...).  I
am wondering if someone could point me towards R packages/functions
appropriate for these data?  I assume that it would evir and I have
read through the documentation, but I would appreciate some guidance
if someone with more experience is willing to offer it.

Many thanks,

Sarah



More information about the R-help mailing list