tune_svm: Train and test Supper Vector Machine classifier.

In anpatton/overdoseR: Predictive Modeling for Opioid Overdoses from EMS Data

Description

Train and test Supper Vector Machine classifier.

Usage

tune_svm(
  training_data,
  testing_data,
  save_model = TRUE,
  model_name = paste0("opioid_svm_model_", Sys.Date()),
  target = "opioid",
  features = "defaults",
  kernel_type = c("radial", "linear"),
  cost_grid = c(0.01, 0.1, 1, 10, 100)
)

Arguments

training_data	formatted EMS data for model training
testing_data	formatted EMS data for model training
save_model	Logical flag to save model as .RDS object
model_name	Filepath and name of model to save, .RDS extension not needed
target	Binary (0/1) for class labels
features	Character vector of features, "default" for recommended options
kernel_type	Radial or linear kernel for SVM
cost_grid	Vector of cost parameters to tune over

Value

List object with model and test data with predictions

• model - Fitted GLM object

• results - Test data with predictions

• tune_object - SVM tuning object

Examples

## Not run: tune_svm(training_data = train, testing_data = test)

anpatton/overdoseR documentation built on July 16, 2020, 1:53 a.m.