glmnetPredict {bestglm} | R Documentation |
Predict by averaging the predictions from cv.glmnet().
glmnetPredict(XyList, NREP = 15, alpha = 0, nfolds=10, family = c("gaussian", "binomial", "poisson", "multinomial"))
XyList |
list with components XyTr, XTr, yTr, XTe. |
NREP |
number of replications to use in average |
alpha |
elastic net parameter |
nfolds |
Number of folds, K, in regularized K-fold CV, must be >3 and <=10. |
family |
model |
vector with predictions
A. I. McLeod
trainTestPartition
,
glmnetGridTable
,
glmnet
,
cv.glmnet
,
predict.glmnet
set.seed(7733551) out <- trainTestPartition(mcdonald) round(glmnetGridTable(out),4) yh <- glmnetPredict(out, NREP=5) sqrt(mean((out$yTe - yh)^2))