Hi all -
I'm trying to find a way to create dummy variables from factors in a
regression. I have been using biglm along the lines of
ff <- log(Price) ~ factor(Colour):factor(Store) +
factor(DummyVar):factor(Colour):factor(Store)
lm1 <- biglm(ff, data=my.dataset)
but because there are lots of colours (>100) and lots of stores
(>250), I run it to memory problems. Now, not every store sells every
colour and so it should be possible to create the matrix of factor
variables myself and greatly reduce the size of the problem. it seems
that lm / biglm use all combinations of factor levels when used in
factor(Colour):factor(Store) so by creating my own matrix of factor
variables i should be able to reduce the size of the problem
considerably.
If i have a data frame
>my.dataset <- data.frame(Price=1:12, Colour= c('red','blue','green'),
Store=c('a', 'b', 'c', 'a', 'c', 'd', 'e', 'e', 'e', 'e', 'b', 'e'),
DummyVar = sort(rep(c(0,1),6)) )
i want to create a data frame with the dummy vars that looks like
red:a red:e blue:b blue:c blue:e green:c green:d green:e
1 0 0 0 0 0 0 0
0 0 1 0 0 0 0 0
0 0 0 0 0 1 0 0
1 0 0 0 0 0 0 0
0 0 0 1 0 0 0 0
0 0 0 0 0 0 1 0
0 1 0 0 0 0 0 0
0 0 0 0 1 0 0 0
0 0 0 0 0 0 0 1
0 1 0 0 0 0 0 0
0 0 1 0 0 0 0 0
0 0 0 0 0 0 0 1
any ideas would be appreciated.
--
Tim Calkins
0406 753 997
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
