Predict for a new dataset by using a simple ensemble. Predictions from individual
models are combined according to fun
Arguments
- object
an simple_ensemble object
- class_thresh
probability threshold used to convert probabilities into classes. It can be a number (between 0 and 1), or a character metric (currently "tss_max", "kap_max" or "sensitivity"). For sensitivity, an additional target value is passed along as a second element of a vector, e.g. c("sensitivity",0.8).
- metric_thresh
a vector of length 2 giving a metric and its threshold, which will be used to prune which models in the ensemble will be used for the prediction. The 'metrics' need to have been computed when the workflow was tuned. The metric's threshold needs to match the value used during prediction. Examples are c("accuracy",0.8) or c("boyce_cont",0.7).
Value
a simple_ensemble object
Examples
test_ens <- simple_ensemble() %>%
add_member(two_class_res[1:3, ], metric = "roc_auc")
test_ens <- calib_class_thresh(test_ens, class_thresh = "tss_max")
test_ens <- calib_class_thresh(test_ens, class_thresh = "kap_max")
test_ens <- calib_class_thresh(test_ens, class_thresh = c("sens", 0.9))