[R-SIG-Finance] Using blockMaxima in fExtremes package for preprocessing the Data

[Ashish Kumar]

Hello, I am trying to find how blockMaxima works ? I run this on danish data
and i get the following :-

> summary(danish)
   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
  1.000   1.321   1.778   3.385   2.967 263.300 

> blockMaxima(danish, block=500)
       82       472       388       356       121 
263.25037  57.41064  32.46753 152.41321 144.65759 
> 

I guess there are 5 blocks of following sizes
       82       472       388       356       121 
with the maximum value of each block being..
263.25037  57.41064  32.46753 152.41321 144.65759 

But if you add all these block i.e. 
      82     +  472  +    388  +    356  +     121  = 1419..

> length(danish)
[1] 2167

The sample size of danish is 2167 which is not equal to 1419...
 
Can someone explain how blockMaxima is preprocessing the tails of danish
data..?


Ashish

-- 
View this message in context: http://www.nabble.com/Using-blockMaxima-in-fExtremes-package-for-preprocessing-the-Data-tf4521138.html#a12897457
Sent from the Rmetrics mailing list archive at Nabble.com.