Has anyone any experience of applying observational weights in bagging? I am
performing regression trees (continuous data on bird abundance) and need to
account for sampling intensity. In a single tree, i.e. a call of rpart, I can
specify weights either by having a separate vector called weights, or by a
variable called weights in the dataframe under analysis. Both produce sensible
(and identical) output. These results are different to an rpart call without a
weights argument and appears to be working as expected.
However, if I analyse these data with a bagging call with a weights argument the
model runs instantaneously and the output produced is empty; no error messages
or warnings are generated. This happens whether the weights are within the
dataframe or in a separate vector. Even assigning the weights in the bagging
call (weights <-1) produces empty output. If I run the data without the
weights argument I get sensible output.
Despite being possible in rpart, are case weights not possible in bagging? I
know that weights are not possible in regression under randomforests.
#this works
r1w <- rpart(how_many ~ covariates, data=ttv.data.sk, weights,
method="poisson")
#this works
b1 <- bagging(how_many ~ covariates, data=ttv.data.sk,
method="poisson", coob=T, nbagg=30)
#this fails
b1w <- bagging(how_many ~ covariates, data=ttv.data.sk, weights,
method="poisson", coob=T, nbagg=30)
thanks in advance
Simon
____________________________________________________________
Sign-up for Bird Atlas 2007-11 at www.birdatlas.net
____________________________________________________________
Dr Simon Gillings
Senior Research Ecologist - Land Use
British Trust for Ornithology
The Nunnery, Thetford, Norfolk, IP24 2PU, UK
Tel +44(0)1842 750050 Fax +44(0)1842 750030
Charity No 216652 (England and Wales)
Company Limited by Guarantee No 357284 (England and Wales)
Registered Office The Nunnery, Thetford, Norfolk IP24 2PU
[[alternative HTML version deleted]]
