[R] Inverse Gaussian distribution not working in glm funciton

michael tsagris mt@@gri@ @ending from y@hoo@gr
Mon Apr 16 09:46:41 CEST 2018


Hello everybody,
I have encountered a problem with the inverse Gaussian distribution. It is very likely that it will not work regardless of the data input. I have programmed this regression and it works fine no matter which distribution the response comes from.
If you run this example (first tried and already got the error)
set.seed(1234)
y = abs( rnorm(150) )
glm( y ~ ., iris, family = inverse.gaussian(log) )
You will get this message: 

Error: inner loop 1; cannot correct step size
In addition: Warning message:
step size truncated due to divergence 
 Whom should I talk to about fixing this bug, in some next update of R?
Michail

	[[alternative HTML version deleted]]




More information about the R-help mailing list