[R] rda.cv cross validation
Alexandre Serra Barreto
fciclone at bol.com.br
Sat Jan 23 11:57:40 CET 2010
Dear RÂ listmates,
I am doing a cross validation with rda.cv to set alpha and delta parameters
in a training set counting on 614 instances. The result of the process in R
is:
rda.cv(fit = z, x = Tt, y = TtL, nfold = 10)
$nonzero
     delta
alpha 0 0.333 0.667 1 1.333 1.667 2 2.333 2.667 3
 0     9    9    9 8    8    8 8   Â
7Â Â Â Â 6 4
 0.11 9    9    8 8    6    3 1    1    1 0
 0.22 9    8    7 6    2    1 1    0    0 0
 0.33 9    8    6 2    1    1 0    0    0 0
 0.44 9    8    4 2    1    0 0    0    0 0
 0.55 9    6    2 1    0    0 0    0    0 0
 0.66 9    6    2 1    0    0 0    0    0 0
 0.77 9    6    2 1    0    0 0    0    0 0
 0.88 9    5    2 0    0    0 0    0    0 0
 0.99 9    4    1 0    0    0 0    0    0 0
$cv.err
     delta
alpha  0 0.333 0.667  1 1.333 1.667  2 2.333 2.667  3
 0     22   23   23 25   25   25 23   Â
25Â Â Â Â 26Â 34
 0.11 24   24   25 28   31    54 64    72  Â
160 214
 0.22 24   25   29 39   58    72 144   214  214
214
 0.33 24   27   36 56   72   160 214   214  214
214
 0.44 25   30   53 64  126   214 214  214  214 214
 0.55 25   32   53 73  214   214 214  214  214 214
 0.66 25   35   57 127  214  214 214  214  214 214
 0.77 25   37   64 201  214  214 214  214  214 214
 0.88 25   42   76 214  214  214 214  214  214 214
 0.99 25   49   85 214  214  214 214  214  214 214
My doubt is: What do these numbers in the second table really represent? I
am considerind that they are the averaged number of missclassifications for
each set. Am I right?
Thanks,
Alexandre.
More information about the R-help
mailing list