validateExternal: validateExternal - Validate model performance on new data

In OHDSI/PatientLevelPrediction: Develop Clinical Prediction Models Using the Common Data Model

validateExternal - Validate model performance on new data

Description

validateExternal - Validate model performance on new data

Usage

validateExternal(
  validationDesignList,
  databaseDetails,
  logSettings = createLogSettings(verbosity = "INFO", logName = "validatePLP"),
  outputFolder,
  cohortDefinitions = NULL
)

Arguments

validationDesignList	A list of objects created with createValidationDesign
databaseDetails	A list of objects of class databaseDetails created using createDatabaseDetails
logSettings	An object of logSettings created using createLogSettings
outputFolder	The directory to save the validation results to
cohortDefinitions	A cohortDefinitionSet object created with CohortGenerator (subfolders are created per database in validationDatabaseDetails)

Value

A list of results

Examples

 
data("simulationProfile")
plpData <- simulatePlpData(simulationProfile, n=1000)
# first fit a model on some data, default is a L1 logistic regression
saveLoc <- file.path(tempdir(), "development")
results <- runPlp(plpData, saveDirectory = saveLoc)
# then create my validation design
validationDesign <- createValidationDesign(1, 3, plpModelList = list(results$model))
# I will validate on Eunomia example database
connectionDetails <- Eunomia::getEunomiaConnectionDetails()
Eunomia::createCohorts(connectionDetails)
databaseDetails <- createDatabaseDetails(connectionDetails = connectionDetails,
cdmDatabaseSchema = "main", cdmDatabaseName = "Eunomia", cdmDatabaseId = 1,
targetId = 1, outcomeIds = 3)
path <- file.path(tempdir(), "validation")
validateExternal(validationDesign, databaseDetails, outputFolder = path)
# see generated result files
dir(path, recursive = TRUE)
# clean up
unlink(saveLoc, recursive = TRUE)
unlink(path, recursive = TRUE)

OHDSI/PatientLevelPrediction documentation built on June 9, 2025, 10:58 a.m.