Hello,
I want to used stratifid kfold cross-validation in brms. It said that ‘group’ should be provided with folds=‘stratified’, but I don’t know what is the group meaning.
I have defined group<-factor(x=fit1$data$y,levels=c(0,1))
but the error is Group ‘011010110111010’ is not a valid grouping factor. Valid groups are: ‘’
I am not sure what should be put in. Could you please help me ?
Thanks!
Looking from K-Fold Cross-Validation — kfold.brmsfit • brms
group : Optional name of a grouping variable or factor in the model.
Can you show your brms model and how you call kfold?
Yes.
#model1
#prior statement
prior1<-c(prior(normal(0,10000),class=b,coef=c(f1,f2,f3,f4,f5,f6)),
prior(normal(0,10000),class=Intercept))
#model statement
model1<-brm(y~f1+f2+f3+f4+f5+f6,
data=s1,family=bernoulli,warmup=5000,iter=20000,chains=1)
g<-factor(x=model1$data$y,levels=c(0,1))
kf1<-kfold(model1,K=5,save_fits=TRUE,chains=1,group=“g”,folds=‘stratified’,compare = TRUE)
I specifed a logistic model and it must require a grouping factor…Thanks!