I tried your gbm.step function to model continuous data in binomial distribution (e.g. proportion of successful cases). This is easy to run in original gbm but I need your additional function output. As the matter of fact your code restricts to binary input and the function fails with my continuous data. It would be really nice, if you could change the input to the full data range [0,1]. I have analysed your code and it would mean a change of code within two lines.
change request
Dear developers,
I tried your gbm.step function to model continuous data in binomial distribution (e.g. proportion of successful cases). This is easy to run in original gbm but I need your additional function output. As the matter of fact your code restricts to binary input and the function fails with my continuous data. It would be really nice, if you could change the input to the full data range [0,1]. I have analysed your code and it would mean a change of code within two lines.
Cheers Anne