Problem running binary trait
vunguyen1907 opened this issue · comments
Hi YiLiLin.
Sorry to bother you again.
I am tring to run binary (survival) trait code as 0, 1 and NA in KAML but can not finish the run. It run well with continuous trait. Please see the screenshot.
Do my phenotype data format is wrong?
I ran this:
iter1test1 <- KAML(pfile="pheno_iter1test1.txt", pheno=1, dcovfile="dcov_tracatfish.txt", qcovfile="qcov_tracatfish.txt", gfile="tracatfish", sample.num=20, crv.num=5)
Regards,
Vu
How is the proportion of case and control at this trait? As KAML uses cross-validation to optimize model parameters, if the distribution of case and control is too extreme, there maybe a risk that all the cases or controls would be allocated into trainings, therefore, all the validations only have one category, which would cause problem to compute prediction accuracy.
@bigshrimp123 Thank for your feedback. Please try to change another seed by parameter 'theSeed', and also turn on the flag binary = TRUE