My name is Ilya Yalchyk, I'm a master student in the University of Bonn, and a research assistant in Fraunhofer SCAI. I'm using your library for our research for Alzheimer's disease.
I've noticed a problem that made it hard for me to use your library. Let's consider an example with a random effect:
library(MASS)
house.mblogit <- mblogit(Sat ~ Infl + Cont, weights = Freq, data = housing, random = ~1|Type)
It converges.
But if we change the way how we provide the formula:
house.mblogit <- mblogit(as.formula("Sat ~ Infl + Cont"), weights = Freq, data = housing, random = ~1|Type)
It produces the following error:
Error in terms.formula(tmp, simplify = TRUE): invalid model formula in ExtractVars
If we provide the formula as a variable:
f <- Sat ~ Infl + Cont
house.mblogit <- mblogit(f, weights = Freq, data = housing, random = ~1|Type)
It produces another error:
Error in attributes(.Data) <- c(attributes(.Data), attrib): cannot set attribute on a symbol
In this pull request, I suggest a simple fix for this problem. I hope it will be useful.
melff
commented
3 years ago
Dear Prof. Dr. Martin Elff,
My name is Ilya Yalchyk, I'm a master student in the University of Bonn, and a research assistant in Fraunhofer SCAI. I'm using your library for our research for Alzheimer's disease.
I've noticed a problem that made it hard for me to use your library. Let's consider an example with a random effect:
It converges.
But if we change the way how we provide the formula:
It produces the following error:
If we provide the formula as a variable:
It produces another error:
In this pull request, I suggest a simple fix for this problem. I hope it will be useful.
Yours sincerely, Ilya Yalchyk