[R] glm() problem
svante
svante at math.aau.dk
Sat Apr 7 20:37:26 CEST 2007
I have some - for me - unresolvable troubles with glm.
First call:
glm(n~.,data=mDat,family=poisson,control = list(trace = TRUE))
Error in 1:control$maxit : NA/NaN argument
Second call:
glm(n~.,data=mDat,family=poisson,control = list(trace = TRUE,maxit=1))
Deviance = 23.48436 Iterations - 1
Error in if (abs(dev - devold)/(0.1 + abs(dev)) < control$epsilon) { :
argument is of length zero
Some of the response counts are zero - so third call:
glm(n~.,data=mDat[mDat$n>0,],family=poisson,control = list(trace =
TRUE,maxit=1))
Deviance = 17.63633 Iterations - 1
Error in if (abs(dev - devold)/(0.1 + abs(dev)) < control$epsilon) { :
argument is of length zero
Anyone having a clue to this problem?
Best wishes
svante
More information about the R-help
mailing list