Dear list,
I have read several posts on this topic. I would use the same
methodology as proposed
by Simon Wood in this post:
http://r.789695.n4.nabble.com/variance-explained-by-each-term-in-a-GAM-td836513.html
My first question is:
Does anyone know a scientific source (paper, book,...) that explains
or uses this
methodology. I have read several articles, particularly in the field of
ecology, that use GAM and I've never read one that dealt with the
proportionof variance explained by each term of a model... and
I think it's an important topic, at least for my study!
My second question:
if I supplied an offset to gam as an argument, should I use it in the full,
reduced and null models, like I did in this example adapted from the
previous post from Simon Wood. I guess so...
## fit full and reduced models...
b <- gam(y~offset(x)+s(x1)+s(x2))
b1 <- gam(y~offset(x)+s(x1),sp=b$sp[
1])
b2 <- gam(y~offset(x)+s(x2),sp=b$sp[2])
b0 <- gam(y~offset(x)+1)
## calculate proportions deviance explained...
dev_x2 <-(deviance(b1)-deviance(b))/deviance(b0) ## prop explained by s(x2)
dev_x1 <-(deviance(b2)-deviance(b))/deviance(b0) ## prop explained by s(x1)
for example, if dev_x2 = 0.4 and dev_x1 =0.1, is the remaining 0.5
corresponds to the proportion of variance explained by the interaction
between the two terms?
Thanks in advance,
Sam
[[alternative HTML version deleted]]
