make_predictions: Make predictions using a model Generic function for testing a...

In crtests: Classification and Regression Tests

Description

Make predictions using a model

Generic function for testing a model by making predictions

Usage

make_predictions(model, data, test, ...)

## Default S3 method:
make_predictions(model, data, test, ...)

## S3 method for class 'rpart'
make_predictions(model, data, test, ...)

## S3 method for class 'boosting'
make_predictions(model, data, test, ...)

## S3 method for class 'gbm'
make_predictions(model, data, test, ...)

Arguments

model	A classification or regression model
data	The list of train and holdout data sets
test	The test being conducted
...	Extra arguments to make_predictions

Methods (by class)

• default: This function is a simple wrapper to predict, which it with the trained model and holdout data. Model classes that require extra arguments to predict can do so through a separate implementations or, less desirably, through the extra arguments.

• rpart: Calls predict.rpart with appropriate type: "class" for classifcation problems and "vector" for regression problems. Other problem types are not supported, providing a test with another class throws an error.

• boosting: Calls predict.boosting on the created model

• gbm: Calls predict.gbm on the created model with n.trees = 100

See Also

predict.rpart

crtests documentation built on May 1, 2019, 9:09 p.m.