Is there any way of identifying the number and position of the knots when using gam in the mgcv package in r? So I have something like this:
fit.add <- gam( rep.pos ~ 0 + factor(AY) + s(dur), family=quasipoisson(link='log')
, data=incur.agg)
And I want to find where this has placed the knots so I can use them in here
dur.knots <- c(6,10)
dur.df <- length(dur.knots) + 1
dur.spline <- ns( dur.start:max.dur, knots=dur.knots )
and save a lot of time trial and error-ing with how many knots and where they should be to give the best fit.
Thanks for the help.
