predict.hhcartr: predict - Create generic S3method to make predictions via...

In hhcartr: HHCART(G) - A Reflected Feature Space for CART

Description

This function creates a generic S3method predict which is used to call predict.hhcartr when an object of type hhcartr passed to the predict function, i.e. an object that is returned from the fit() function. The object created from the predict function supports the accuracy and predictions methods. The accuracy method returns the accuracy achieved on the test_data and the method predictions returns the actual predictions made on the test_data.

Usage

## S3 method for class 'hhcartr'
predict(object, ..., test_data)

Arguments

object	Unused parameter.
...	Unused parameter.
test_data	The test dataset the user wants to make predictions on.

Value

exposes the accuracy() and predictions() methods.

Examples

# source: /man/examples/predict.R

# Basic usage of predict().

# Note: we need to have a model to modify first.

# load our data.
X <- iris[,1:4]
y <- iris[,5]

# instantiate our model.
clf = HHDecisionTree(n_folds=1,
                     n_trees=1,
                     pruning=FALSE,
                     min_node_impurity=0.0)

# describe what dataset our model is using.
setDataDescription("IRIS Dataset")

# train our model.
model_output <- clf$fit(X, y)

# make predictions on an unseen test set.
# As the IRIS dataset provides no separate test set,
# for the sake of this example we will predict on the original data
# and pretend it is previously unseen data.
preds <- predict(model_output, test_data = iris)

# The predict object 'preds' exposes the following methods
# that we can use to extract the results of interest
# (depending upon model options used):

# preds$accuracy and preds$predictions.

hhcartr documentation built on July 2, 2021, 9:06 a.m.