Skip to contents

Predict from an idiographic ML result

Usage

# S3 method for class 'idioml_result'
predict(
  object,
  newdata = NULL,
  scope = c("pooled", "individual"),
  model = NULL,
  estimator = NULL,
  type = c("response", "class"),
  ...
)

Arguments

object

An idioml_result.

newdata

Optional new data. If NULL, the stored test-set predictions are returned.

scope

"pooled" or "individual".

model

Fitted model name to use. Default is the first fitted model.

estimator

Fitted estimator/backend for model. Default is that model's first fitted estimator.

type

"response" for numeric predictions/probabilities or "class" for classification labels.

...

Ignored.

Value

A data.frame of predictions. With newdata = NULL, the stored held-out predictions of the chosen scope, model and estimator, one row per test occasion; otherwise one row per row of newdata.

Examples

# Predicting new rows needs the fitted models, so keep them.
fit <- fit_ml(srl, "effort", c("efficacy", "planning"), "name",
              model = "ridge", keep_fits = TRUE)
predict(fit, newdata = head(srl), scope = "pooled")
#>   model_scope model estimator subject original_row predicted
#> 1      pooled ridge    native   Aisha            1  35.49444
#> 2      pooled ridge    native   Aisha            2  44.40581
#> 3      pooled ridge    native   Aisha            3  61.59341
#> 4      pooled ridge    native   Aisha            4  62.36215
#> 5      pooled ridge    native   Aisha            5  57.93781
#> 6      pooled ridge    native   Aisha            6  54.28221