[R] stemplot of data... now the distribution....

nelson - nelson1977 at gmail.com
Fri Oct 27 15:59:52 CEST 2006


Hi all,
  i have a dataset that have this stemplot

  0 | 11111111111111111111111111111111111111111122222222233333333333334444
  0 | 5555556666666667777777788889999999999
  1 | 000000111111112222222233333344444444
  1 | 55555555555566666677888899999999
  2 | 0000011123344
  2 | 5556666666778889999
  3 | 0000112234
  3 | 556666777889999
  4 | 000000002233334
  4 | 55555777788899
  5 | 01111111222222233333444444444
  5 | 555555555666666777777777888899999
  6 | 00000011112222333444
  6 | 55556677788899
  7 | 0012223334
  7 | 77
  8 | 4
  8 | 5
  9 | 0


it seems that the data have 2 modes (it's true?). So can i use the mix
package to fit it using a mixture of distributions? (i think a
exponential, a gumbel for the first part and a normal with mean on the
secondo mode... it's a right to do it? Or i can think to have data
distribudet following a weibull and a normal.... how can i test what
distro fits well? i think to use kolmogorov smirnov test and
chi-square test, that's ok?

Thanks,

   nelson



More information about the R-help mailing list