hanase / BMA

R package for Bayesian model averaging
34 stars 11 forks source link

Bug in back-transforming variable names in bic.glm #1

Open hanase opened 7 years ago

hanase commented 7 years ago

It only appears if one of the variable names has a number at the end and there is a factor involved. In such a case the function adds a suffix ".x" to the variable names in order not to confuse the normal variables (with numbers at the end) with factors (see here)

However the back-transformation of the names of factors is not correct, especially for names of factor varaibles, since their name is e.g. myvar.x.1. See example of Gomaa.