R/decisionpanel.h.R
In sbalci/ClinicoPathJamoviModule: Analysis for Clinicopathological Research
Defines functions
decisionpanel
Documented in
decisionpanel
# This file is automatically generated, you probably don't want to edit this
decisionpanelOptions <- if (requireNamespace("jmvcore", quietly=TRUE)) R6::R6Class(
"decisionpanelOptions",
inherit = jmvcore::Options,
public = list(
initialize = function(
tests = NULL,
testLevels = "",
gold = NULL,
goldPositive = NULL,
useCosts = FALSE,
testCosts = "",
fpCost = 100,
fnCost = 1000,
strategies = "all",
parallelRules = "any",
customThreshold = 2,
maxTests = 3,
sequentialStop = "positive",
optimizationCriteria = "accuracy",
minSensitivity = 0.8,
minSpecificity = 0.8,
createTree = FALSE,
treeMethod = "costSensitive",
maxDepth = 5,
minSplit = 20,
showAllCombinations = FALSE,
topN = 10,
compareStrategies = TRUE,
bootstrap = FALSE,
bootReps = 1000,
plotTree = TRUE,
plotComparison = TRUE,
plotCostEffect = TRUE,
plotROC = FALSE,
prevalence = 0,
crossValidate = FALSE,
nFolds = 5,
seed = 12345, ...) {
super$initialize(
package="ClinicoPath",
name="decisionpanel",
requiresData=TRUE,
...)
private$..tests <- jmvcore::OptionVariables$new(
"tests",
tests,
suggested=list(
"nominal"),
permitted=list(
"factor"),
rejectUnusedLevels=TRUE)
private$..testLevels <- jmvcore::OptionString$new(
"testLevels",
testLevels,
default="")
private$..gold <- jmvcore::OptionVariable$new(
"gold",
gold,
suggested=list(
"nominal"),
permitted=list(
"factor"))
private$..goldPositive <- jmvcore::OptionLevel$new(
"goldPositive",
goldPositive,
variable="(gold)")
private$..useCosts <- jmvcore::OptionBool$new(
"useCosts",
useCosts,
default=FALSE)
private$..testCosts <- jmvcore::OptionString$new(
"testCosts",
testCosts,
default="")
private$..fpCost <- jmvcore::OptionNumber$new(
"fpCost",
fpCost,
default=100,
min=0)
private$..fnCost <- jmvcore::OptionNumber$new(
"fnCost",
fnCost,
default=1000,
min=0)
private$..strategies <- jmvcore::OptionList$new(
"strategies",
strategies,
options=list(
"all",
"single",
"parallel",
"sequential",
"custom"),
default="all")
private$..parallelRules <- jmvcore::OptionList$new(
"parallelRules",
parallelRules,
options=list(
"any",
"all",
"majority",
"custom"),
default="any")
private$..customThreshold <- jmvcore::OptionNumber$new(
"customThreshold",
customThreshold,
default=2,
min=1)
private$..maxTests <- jmvcore::OptionNumber$new(
"maxTests",
maxTests,
default=3,
min=2,
max=10)
private$..sequentialStop <- jmvcore::OptionList$new(
"sequentialStop",
sequentialStop,
options=list(
"positive",
"negative",
"confirmatory",
"exclusion"),
default="positive")
private$..optimizationCriteria <- jmvcore::OptionList$new(
"optimizationCriteria",
optimizationCriteria,
options=list(
"accuracy",
"sensitivity",
"specificity",
"ppv",
"npv",
"youden",
"utility",
"efficiency"),
default="accuracy")
private$..minSensitivity <- jmvcore::OptionNumber$new(
"minSensitivity",
minSensitivity,
default=0.8,
min=0,
max=1)
private$..minSpecificity <- jmvcore::OptionNumber$new(
"minSpecificity",
minSpecificity,
default=0.8,
min=0,
max=1)
private$..createTree <- jmvcore::OptionBool$new(
"createTree",
createTree,
default=FALSE)
private$..treeMethod <- jmvcore::OptionList$new(
"treeMethod",
treeMethod,
options=list(
"cart",
"conditional",
"costSensitive",
"ensemble"),
default="costSensitive")
private$..maxDepth <- jmvcore::OptionNumber$new(
"maxDepth",
maxDepth,
default=5,
min=1,
max=10)
private$..minSplit <- jmvcore::OptionNumber$new(
"minSplit",
minSplit,
default=20,
min=5)
private$..showAllCombinations <- jmvcore::OptionBool$new(
"showAllCombinations",
showAllCombinations,
default=FALSE)
private$..topN <- jmvcore::OptionNumber$new(
"topN",
topN,
default=10,
min=1,
max=50)
private$..compareStrategies <- jmvcore::OptionBool$new(
"compareStrategies",
compareStrategies,
default=TRUE)
private$..bootstrap <- jmvcore::OptionBool$new(
"bootstrap",
bootstrap,
default=FALSE)
private$..bootReps <- jmvcore::OptionNumber$new(
"bootReps",
bootReps,
default=1000,
min=100,
max=10000)
private$..plotTree <- jmvcore::OptionBool$new(
"plotTree",
plotTree,
default=TRUE)
private$..plotComparison <- jmvcore::OptionBool$new(
"plotComparison",
plotComparison,
default=TRUE)
private$..plotCostEffect <- jmvcore::OptionBool$new(
"plotCostEffect",
plotCostEffect,
default=TRUE)
private$..plotROC <- jmvcore::OptionBool$new(
"plotROC",
plotROC,
default=FALSE)
private$..prevalence <- jmvcore::OptionNumber$new(
"prevalence",
prevalence,
default=0,
min=0,
max=1)
private$..crossValidate <- jmvcore::OptionBool$new(
"crossValidate",
crossValidate,
default=FALSE)
private$..nFolds <- jmvcore::OptionNumber$new(
"nFolds",
nFolds,
default=5,
min=2,
max=10)
private$..seed <- jmvcore::OptionNumber$new(
"seed",
seed,
default=12345)
self$.addOption(private$..tests)
self$.addOption(private$..testLevels)
self$.addOption(private$..gold)
self$.addOption(private$..goldPositive)
self$.addOption(private$..useCosts)
self$.addOption(private$..testCosts)
self$.addOption(private$..fpCost)
self$.addOption(private$..fnCost)
self$.addOption(private$..strategies)
self$.addOption(private$..parallelRules)
self$.addOption(private$..customThreshold)
self$.addOption(private$..maxTests)
self$.addOption(private$..sequentialStop)
self$.addOption(private$..optimizationCriteria)
self$.addOption(private$..minSensitivity)
self$.addOption(private$..minSpecificity)
self$.addOption(private$..createTree)
self$.addOption(private$..treeMethod)
self$.addOption(private$..maxDepth)
self$.addOption(private$..minSplit)
self$.addOption(private$..showAllCombinations)
self$.addOption(private$..topN)
self$.addOption(private$..compareStrategies)
self$.addOption(private$..bootstrap)
self$.addOption(private$..bootReps)
self$.addOption(private$..plotTree)
self$.addOption(private$..plotComparison)
self$.addOption(private$..plotCostEffect)
self$.addOption(private$..plotROC)
self$.addOption(private$..prevalence)
self$.addOption(private$..crossValidate)
self$.addOption(private$..nFolds)
self$.addOption(private$..seed)
}),
active = list(
tests = function() private$..tests$value,
testLevels = function() private$..testLevels$value,
gold = function() private$..gold$value,
goldPositive = function() private$..goldPositive$value,
useCosts = function() private$..useCosts$value,
testCosts = function() private$..testCosts$value,
fpCost = function() private$..fpCost$value,
fnCost = function() private$..fnCost$value,
strategies = function() private$..strategies$value,
parallelRules = function() private$..parallelRules$value,
customThreshold = function() private$..customThreshold$value,
maxTests = function() private$..maxTests$value,
sequentialStop = function() private$..sequentialStop$value,
optimizationCriteria = function() private$..optimizationCriteria$value,
minSensitivity = function() private$..minSensitivity$value,
minSpecificity = function() private$..minSpecificity$value,
createTree = function() private$..createTree$value,
treeMethod = function() private$..treeMethod$value,
maxDepth = function() private$..maxDepth$value,
minSplit = function() private$..minSplit$value,
showAllCombinations = function() private$..showAllCombinations$value,
topN = function() private$..topN$value,
compareStrategies = function() private$..compareStrategies$value,
bootstrap = function() private$..bootstrap$value,
bootReps = function() private$..bootReps$value,
plotTree = function() private$..plotTree$value,
plotComparison = function() private$..plotComparison$value,
plotCostEffect = function() private$..plotCostEffect$value,
plotROC = function() private$..plotROC$value,
prevalence = function() private$..prevalence$value,
crossValidate = function() private$..crossValidate$value,
nFolds = function() private$..nFolds$value,
seed = function() private$..seed$value),
private = list(
..tests = NA,
..testLevels = NA,
..gold = NA,
..goldPositive = NA,
..useCosts = NA,
..testCosts = NA,
..fpCost = NA,
..fnCost = NA,
..strategies = NA,
..parallelRules = NA,
..customThreshold = NA,
..maxTests = NA,
..sequentialStop = NA,
..optimizationCriteria = NA,
..minSensitivity = NA,
..minSpecificity = NA,
..createTree = NA,
..treeMethod = NA,
..maxDepth = NA,
..minSplit = NA,
..showAllCombinations = NA,
..topN = NA,
..compareStrategies = NA,
..bootstrap = NA,
..bootReps = NA,
..plotTree = NA,
..plotComparison = NA,
..plotCostEffect = NA,
..plotROC = NA,
..prevalence = NA,
..crossValidate = NA,
..nFolds = NA,
..seed = NA)
)
decisionpanelResults <- if (requireNamespace("jmvcore", quietly=TRUE)) R6::R6Class(
"decisionpanelResults",
inherit = jmvcore::Group,
active = list(
summary = function() private$.items[["summary"]],
optimalPanel = function() private$.items[["optimalPanel"]],
strategyComparison = function() private$.items[["strategyComparison"]],
individualTests = function() private$.items[["individualTests"]],
allCombinations = function() private$.items[["allCombinations"]],
treeStructure = function() private$.items[["treeStructure"]],
treePerformance = function() private$.items[["treePerformance"]],
crossValidation = function() private$.items[["crossValidation"]],
bootstrapResults = function() private$.items[["bootstrapResults"]],
treeVisualization = function() private$.items[["treeVisualization"]],
strategyComparisonPlot = function() private$.items[["strategyComparisonPlot"]],
costEffectivenessPlot = function() private$.items[["costEffectivenessPlot"]],
rocCurvesPlot = function() private$.items[["rocCurvesPlot"]],
recommendations = function() private$.items[["recommendations"]]),
private = list(),
public=list(
initialize=function(options) {
super$initialize(
options=options,
name="",
title="Decision Panel Optimization",
refs=list(
"rpart",
"partykit",
"DiagrammeR",
"ROCR",
"ClinicoPathJamoviModule"))
self$add(jmvcore::Html$new(
options=options,
name="summary",
title="Summary",
visible=TRUE))
self$add(jmvcore::Table$new(
options=options,
name="optimalPanel",
title="Optimal Test Panel",
visible=TRUE,
columns=list(
list(
`name`="rank",
`title`="Rank",
`type`="integer"),
list(
`name`="tests",
`title`="Test Combination",
`type`="text"),
list(
`name`="strategy",
`title`="Strategy",
`type`="text"),
list(
`name`="sensitivity",
`title`="Sensitivity",
`type`="number",
`format`="pc"),
list(
`name`="specificity",
`title`="Specificity",
`type`="number",
`format`="pc"),
list(
`name`="accuracy",
`title`="Accuracy",
`type`="number",
`format`="pc"),
list(
`name`="ppv",
`title`="PPV",
`type`="number",
`format`="pc"),
list(
`name`="npv",
`title`="NPV",
`type`="number",
`format`="pc"),
list(
`name`="youden",
`title`="Youden's J",
`type`="number",
`format`="zto"),
list(
`name`="cost",
`title`="Cost",
`type`="number",
`format`="zto",
`visible`="(useCosts)"),
list(
`name`="utility",
`title`="Utility",
`type`="number",
`format`="zto",
`visible`="(useCosts)"),
list(
`name`="efficiency",
`title`="Efficiency",
`type`="number",
`format`="zto",
`visible`="(useCosts)"))))
self$add(jmvcore::Table$new(
options=options,
name="strategyComparison",
title="Testing Strategy Comparison",
visible="(compareStrategies)",
columns=list(
list(
`name`="strategy",
`title`="Strategy",
`type`="text"),
list(
`name`="nTests",
`title`="Tests Used",
`type`="integer"),
list(
`name`="avgSens",
`title`="Avg Sensitivity",
`type`="number",
`format`="pc"),
list(
`name`="avgSpec",
`title`="Avg Specificity",
`type`="number",
`format`="pc"),
list(
`name`="avgAccuracy",
`title`="Avg Accuracy",
`type`="number",
`format`="pc"),
list(
`name`="avgCost",
`title`="Avg Cost",
`type`="number",
`format`="zto",
`visible`="(useCosts)"),
list(
`name`="bestPanel",
`title`="Best Panel",
`type`="text"),
list(
`name`="bestPerformance",
`title`="Best Performance",
`type`="number",
`format`="zto"))))
self$add(jmvcore::Table$new(
options=options,
name="individualTests",
title="Individual Test Performance",
visible=TRUE,
columns=list(
list(
`name`="test",
`title`="Test",
`type`="text"),
list(
`name`="sensitivity",
`title`="Sensitivity",
`type`="number",
`format`="pc"),
list(
`name`="specificity",
`title`="Specificity",
`type`="number",
`format`="pc"),
list(
`name`="accuracy",
`title`="Accuracy",
`type`="number",
`format`="pc"),
list(
`name`="ppv",
`title`="PPV",
`type`="number",
`format`="pc"),
list(
`name`="npv",
`title`="NPV",
`type`="number",
`format`="pc"),
list(
`name`="plr",
`title`="+LR",
`type`="number",
`format`="zto"),
list(
`name`="nlr",
`title`="-LR",
`type`="number",
`format`="zto"),
list(
`name`="cost",
`title`="Cost",
`type`="number",
`format`="zto",
`visible`="(useCosts)"))))
self$add(jmvcore::Table$new(
options=options,
name="allCombinations",
title="All Test Combinations",
visible="(showAllCombinations)",
columns=list(
list(
`name`="tests",
`title`="Tests",
`type`="text"),
list(
`name`="strategy",
`title`="Strategy",
`type`="text"),
list(
`name`="nTests",
`title`="N Tests",
`type`="integer"),
list(
`name`="sensitivity",
`title`="Sensitivity",
`type`="number",
`format`="pc"),
list(
`name`="specificity",
`title`="Specificity",
`type`="number",
`format`="pc"),
list(
`name`="accuracy",
`title`="Accuracy",
`type`="number",
`format`="pc"),
list(
`name`="ppv",
`title`="PPV",
`type`="number",
`format`="pc"),
list(
`name`="npv",
`title`="NPV",
`type`="number",
`format`="pc"),
list(
`name`="cost",
`title`="Cost",
`type`="number",
`format`="zto",
`visible`="(useCosts)"),
list(
`name`="utility",
`title`="Utility",
`type`="number",
`format`="zto",
`visible`="(useCosts)"))))
self$add(jmvcore::Html$new(
options=options,
name="treeStructure",
title="Decision Tree Structure",
visible="(createTree)"))
self$add(jmvcore::Table$new(
options=options,
name="treePerformance",
title="Decision Tree Performance",
visible="(createTree)",
columns=list(
list(
`name`="node",
`title`="Node",
`type`="text"),
list(
`name`="condition",
`title`="Condition",
`type`="text"),
list(
`name`="nCases",
`title`="N Cases",
`type`="integer"),
list(
`name`="probPositive",
`title`="P(Disease+)",
`type`="number",
`format`="pc"),
list(
`name`="decision",
`title`="Decision",
`type`="text"),
list(
`name`="accuracy",
`title`="Node Accuracy",
`type`="number",
`format`="pc"))))
self$add(jmvcore::Table$new(
options=options,
name="crossValidation",
title="Cross-Validation Results",
visible="(crossValidate)",
columns=list(
list(
`name`="panel",
`title`="Test Panel",
`type`="text"),
list(
`name`="strategy",
`title`="Strategy",
`type`="text"),
list(
`name`="cvSensitivity",
`title`="CV Sensitivity",
`type`="number",
`format`="pc"),
list(
`name`="cvSensitivitySD",
`title`="SD",
`type`="number",
`format`="zto"),
list(
`name`="cvSpecificity",
`title`="CV Specificity",
`type`="number",
`format`="pc"),
list(
`name`="cvSpecificitySD",
`title`="SD",
`type`="number",
`format`="zto"),
list(
`name`="cvAccuracy",
`title`="CV Accuracy",
`type`="number",
`format`="pc"),
list(
`name`="cvAccuracySD",
`title`="SD",
`type`="number",
`format`="zto"))))
self$add(jmvcore::Table$new(
options=options,
name="bootstrapResults",
title="Bootstrap Confidence Intervals",
visible="(bootstrap)",
columns=list(
list(
`name`="panel",
`title`="Test Panel",
`type`="text"),
list(
`name`="metric",
`title`="Metric",
`type`="text"),
list(
`name`="estimate",
`title`="Estimate",
`type`="number",
`format`="zto"),
list(
`name`="lower",
`title`="2.5%",
`type`="number",
`format`="zto"),
list(
`name`="upper",
`title`="97.5%",
`type`="number",
`format`="zto"),
list(
`name`="se",
`title`="SE",
`type`="number",
`format`="zto"))))
self$add(jmvcore::Image$new(
options=options,
name="treeVisualization",
title="Decision Tree",
width=800,
height=600,
renderFun=".plotTree",
visible="(plotTree && createTree)",
clearWith=list(
"tests",
"gold",
"goldPositive",
"treeMethod",
"maxDepth",
"minSplit")))
self$add(jmvcore::Image$new(
options=options,
name="strategyComparisonPlot",
title="Strategy Comparison",
width=700,
height=500,
renderFun=".plotStrategyComparison",
visible="(plotComparison && compareStrategies)",
clearWith=list(
"tests",
"strategies",
"optimizationCriteria")))
self$add(jmvcore::Image$new(
options=options,
name="costEffectivenessPlot",
title="Cost-Effectiveness Frontier",
width=700,
height=500,
renderFun=".plotCostEffectiveness",
visible="(plotCostEffect && useCosts)",
clearWith=list(
"tests",
"testCosts",
"fpCost",
"fnCost")))
self$add(jmvcore::Image$new(
options=options,
name="rocCurvesPlot",
title="ROC Curves for Top Panels",
width=700,
height=500,
renderFun=".plotROCCurves",
visible="(plotROC)",
clearWith=list(
"tests",
"topN")))
self$add(jmvcore::Html$new(
options=options,
name="recommendations",
title="Recommendations",
visible=TRUE))}))
decisionpanelBase <- if (requireNamespace("jmvcore", quietly=TRUE)) R6::R6Class(
"decisionpanelBase",
inherit = jmvcore::Analysis,
public = list(
initialize = function(options, data=NULL, datasetId="", analysisId="", revision=0) {
super$initialize(
package = "ClinicoPath",
name = "decisionpanel",
version = c(0,0,3),
options = options,
results = decisionpanelResults$new(options=options),
data = data,
datasetId = datasetId,
analysisId = analysisId,
revision = revision,
pause = NULL,
completeWhenFilled = FALSE,
requiresMissings = FALSE,
weightsSupport = 'auto')
}))
#' Decision Panel Optimization
#'
#' Optimize diagnostic test panels by evaluating various combination
#' strategies including parallel testing (cotest), sequential testing, and
#' repeated tests. Creates decision trees to minimize cost while maximizing
#' accuracy.
#'
#'
#' @examples
#' \donttest{
#' # example will be added
#'}
#' @param data The data as a data frame.
#' @param tests Variables representing different diagnostic tests. Can select
#' multiple tests (up to 10) for panel optimization.
#' @param testLevels String specifying positive test levels. Can be: - Single
#' value (used for all tests): "Positive" - Multiple values (comma-separated):
#' "Positive,Yes,1" - Empty string (auto-detection): ""
#' @param gold The gold standard variable for disease classification.
#' @param goldPositive The level of the gold standard variable that indicates
#' disease presence.
#' @param useCosts Include cost considerations in the optimization process.
#' @param testCosts Comma-separated costs for each test in the same order as
#' selected tests. Example: "10,25,50" for three tests.
#' @param fpCost Cost or harm associated with a false positive result.
#' @param fnCost Cost or harm associated with a false negative result.
#' @param strategies Which testing strategies to evaluate in the analysis.
#' @param parallelRules Rule for combining results in parallel testing.
#' @param customThreshold Number of positive tests required for overall
#' positive result (when using custom rule).
#' @param maxTests Maximum number of tests to combine in any strategy.
#' @param sequentialStop When to stop testing in sequential strategies.
#' @param optimizationCriteria Primary criterion for optimizing test panels.
#' @param minSensitivity Minimum sensitivity constraint for panel selection.
#' @param minSpecificity Minimum specificity constraint for panel selection.
#' @param createTree Generate an optimal decision tree for test sequencing.
#' @param treeMethod Method for constructing the decision tree.
#' @param maxDepth Maximum depth of the decision tree.
#' @param minSplit Minimum number of observations required to split a node.
#' @param showAllCombinations Display performance metrics for all possible
#' test combinations.
#' @param topN Number of best-performing panels to display in results.
#' @param compareStrategies Show comparative analysis of different testing
#' strategies.
#' @param bootstrap Calculate bootstrap confidence intervals for performance
#' metrics.
#' @param bootReps Number of bootstrap replications for confidence intervals.
#' @param plotTree Display visual representation of the optimal decision tree.
#' @param plotComparison Create comparison plots for different testing
#' strategies.
#' @param plotCostEffect Create cost-effectiveness frontier plot.
#' @param plotROC Display ROC curves for top performing panels.
#' @param prevalence Known disease prevalence (0 = use sample prevalence).
#' @param crossValidate Perform k-fold cross-validation for panel performance.
#' @param nFolds Number of folds for cross-validation.
#' @param seed Random seed for reproducibility in bootstrap and
#' cross-validation.
#' @return A results object containing:
#' \tabular{llllll}{
#' \code{results$summary} \tab \tab \tab \tab \tab a html \cr
#' \code{results$optimalPanel} \tab \tab \tab \tab \tab a table \cr
#' \code{results$strategyComparison} \tab \tab \tab \tab \tab a table \cr
#' \code{results$individualTests} \tab \tab \tab \tab \tab a table \cr
#' \code{results$allCombinations} \tab \tab \tab \tab \tab a table \cr
#' \code{results$treeStructure} \tab \tab \tab \tab \tab a html \cr
#' \code{results$treePerformance} \tab \tab \tab \tab \tab a table \cr
#' \code{results$crossValidation} \tab \tab \tab \tab \tab a table \cr
#' \code{results$bootstrapResults} \tab \tab \tab \tab \tab a table \cr
#' \code{results$treeVisualization} \tab \tab \tab \tab \tab an image \cr
#' \code{results$strategyComparisonPlot} \tab \tab \tab \tab \tab an image \cr
#' \code{results$costEffectivenessPlot} \tab \tab \tab \tab \tab an image \cr
#' \code{results$rocCurvesPlot} \tab \tab \tab \tab \tab an image \cr
#' \code{results$recommendations} \tab \tab \tab \tab \tab a html \cr
#' }
#'
#' Tables can be converted to data frames with \code{asDF} or \code{\link{as.data.frame}}. For example:
#'
#' \code{results$optimalPanel$asDF}
#'
#' \code{as.data.frame(results$optimalPanel)}
#'
#' @export
decisionpanel <- function(
data,
tests,
testLevels = "",
gold,
goldPositive,
useCosts = FALSE,
testCosts = "",
fpCost = 100,
fnCost = 1000,
strategies = "all",
parallelRules = "any",
customThreshold = 2,
maxTests = 3,
sequentialStop = "positive",
optimizationCriteria = "accuracy",
minSensitivity = 0.8,
minSpecificity = 0.8,
createTree = FALSE,
treeMethod = "costSensitive",
maxDepth = 5,
minSplit = 20,
showAllCombinations = FALSE,
topN = 10,
compareStrategies = TRUE,
bootstrap = FALSE,
bootReps = 1000,
plotTree = TRUE,
plotComparison = TRUE,
plotCostEffect = TRUE,
plotROC = FALSE,
prevalence = 0,
crossValidate = FALSE,
nFolds = 5,
seed = 12345) {
if ( ! requireNamespace("jmvcore", quietly=TRUE))
stop("decisionpanel requires jmvcore to be installed (restart may be required)")
if ( ! missing(tests)) tests <- jmvcore::resolveQuo(jmvcore::enquo(tests))
if ( ! missing(gold)) gold <- jmvcore::resolveQuo(jmvcore::enquo(gold))
if (missing(data))
data <- jmvcore::marshalData(
parent.frame(),
`if`( ! missing(tests), tests, NULL),
`if`( ! missing(gold), gold, NULL))
for (v in tests) if (v %in% names(data)) data[[v]] <- as.factor(data[[v]])
for (v in gold) if (v %in% names(data)) data[[v]] <- as.factor(data[[v]])
options <- decisionpanelOptions$new(
tests = tests,
testLevels = testLevels,
gold = gold,
goldPositive = goldPositive,
useCosts = useCosts,
testCosts = testCosts,
fpCost = fpCost,
fnCost = fnCost,
strategies = strategies,
parallelRules = parallelRules,
customThreshold = customThreshold,
maxTests = maxTests,
sequentialStop = sequentialStop,
optimizationCriteria = optimizationCriteria,
minSensitivity = minSensitivity,
minSpecificity = minSpecificity,
createTree = createTree,
treeMethod = treeMethod,
maxDepth = maxDepth,
minSplit = minSplit,
showAllCombinations = showAllCombinations,
topN = topN,
compareStrategies = compareStrategies,
bootstrap = bootstrap,
bootReps = bootReps,
plotTree = plotTree,
plotComparison = plotComparison,
plotCostEffect = plotCostEffect,
plotROC = plotROC,
prevalence = prevalence,
crossValidate = crossValidate,
nFolds = nFolds,
seed = seed)
analysis <- decisionpanelClass$new(
options = options,
data = data)
analysis$run()
analysis$results
}
sbalci/ClinicoPathJamoviModule documentation built on June 13, 2025, 9:34 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions decisionpanel
Documented in decisionpanel
# This file is automatically generated, you probably don't want to edit this
decisionpanelOptions <- if (requireNamespace("jmvcore", quietly=TRUE)) R6::R6Class(
"decisionpanelOptions",
inherit = jmvcore::Options,
public = list(
initialize = function(
tests = NULL,
testLevels = "",
gold = NULL,
goldPositive = NULL,
useCosts = FALSE,
testCosts = "",
fpCost = 100,
fnCost = 1000,
strategies = "all",
parallelRules = "any",
customThreshold = 2,
maxTests = 3,
sequentialStop = "positive",
optimizationCriteria = "accuracy",
minSensitivity = 0.8,
minSpecificity = 0.8,
createTree = FALSE,
treeMethod = "costSensitive",
maxDepth = 5,
minSplit = 20,
showAllCombinations = FALSE,
topN = 10,
compareStrategies = TRUE,
bootstrap = FALSE,
bootReps = 1000,
plotTree = TRUE,
plotComparison = TRUE,
plotCostEffect = TRUE,
plotROC = FALSE,
prevalence = 0,
crossValidate = FALSE,
nFolds = 5,
seed = 12345, ...) {
super$initialize(
package="ClinicoPath",
name="decisionpanel",
requiresData=TRUE,
...)
private$..tests <- jmvcore::OptionVariables$new(
"tests",
tests,
suggested=list(
"nominal"),
permitted=list(
"factor"),
rejectUnusedLevels=TRUE)
private$..testLevels <- jmvcore::OptionString$new(
"testLevels",
testLevels,
default="")
private$..gold <- jmvcore::OptionVariable$new(
"gold",
gold,
suggested=list(
"nominal"),
permitted=list(
"factor"))
private$..goldPositive <- jmvcore::OptionLevel$new(
"goldPositive",
goldPositive,
variable="(gold)")
private$..useCosts <- jmvcore::OptionBool$new(
"useCosts",
useCosts,
default=FALSE)
private$..testCosts <- jmvcore::OptionString$new(
"testCosts",
testCosts,
default="")
private$..fpCost <- jmvcore::OptionNumber$new(
"fpCost",
fpCost,
default=100,
min=0)
private$..fnCost <- jmvcore::OptionNumber$new(
"fnCost",
fnCost,
default=1000,
min=0)
private$..strategies <- jmvcore::OptionList$new(
"strategies",
strategies,
options=list(
"all",
"single",
"parallel",
"sequential",
"custom"),
default="all")
private$..parallelRules <- jmvcore::OptionList$new(
"parallelRules",
parallelRules,
options=list(
"any",
"all",
"majority",
"custom"),
default="any")
private$..customThreshold <- jmvcore::OptionNumber$new(
"customThreshold",
customThreshold,
default=2,
min=1)
private$..maxTests <- jmvcore::OptionNumber$new(
"maxTests",
maxTests,
default=3,
min=2,
max=10)
private$..sequentialStop <- jmvcore::OptionList$new(
"sequentialStop",
sequentialStop,
options=list(
"positive",
"negative",
"confirmatory",
"exclusion"),
default="positive")
private$..optimizationCriteria <- jmvcore::OptionList$new(
"optimizationCriteria",
optimizationCriteria,
options=list(
"accuracy",
"sensitivity",
"specificity",
"ppv",
"npv",
"youden",
"utility",
"efficiency"),
default="accuracy")
private$..minSensitivity <- jmvcore::OptionNumber$new(
"minSensitivity",
minSensitivity,
default=0.8,
min=0,
max=1)
private$..minSpecificity <- jmvcore::OptionNumber$new(
"minSpecificity",
minSpecificity,
default=0.8,
min=0,
max=1)
private$..createTree <- jmvcore::OptionBool$new(
"createTree",
createTree,
default=FALSE)
private$..treeMethod <- jmvcore::OptionList$new(
"treeMethod",
treeMethod,
options=list(
"cart",
"conditional",
"costSensitive",
"ensemble"),
default="costSensitive")
private$..maxDepth <- jmvcore::OptionNumber$new(
"maxDepth",
maxDepth,
default=5,
min=1,
max=10)
private$..minSplit <- jmvcore::OptionNumber$new(
"minSplit",
minSplit,
default=20,
min=5)
private$..showAllCombinations <- jmvcore::OptionBool$new(
"showAllCombinations",
showAllCombinations,
default=FALSE)
private$..topN <- jmvcore::OptionNumber$new(
"topN",
topN,
default=10,
min=1,
max=50)
private$..compareStrategies <- jmvcore::OptionBool$new(
"compareStrategies",
compareStrategies,
default=TRUE)
private$..bootstrap <- jmvcore::OptionBool$new(
"bootstrap",
bootstrap,
default=FALSE)
private$..bootReps <- jmvcore::OptionNumber$new(
"bootReps",
bootReps,
default=1000,
min=100,
max=10000)
private$..plotTree <- jmvcore::OptionBool$new(
"plotTree",
plotTree,
default=TRUE)
private$..plotComparison <- jmvcore::OptionBool$new(
"plotComparison",
plotComparison,
default=TRUE)
private$..plotCostEffect <- jmvcore::OptionBool$new(
"plotCostEffect",
plotCostEffect,
default=TRUE)
private$..plotROC <- jmvcore::OptionBool$new(
"plotROC",
plotROC,
default=FALSE)
private$..prevalence <- jmvcore::OptionNumber$new(
"prevalence",
prevalence,
default=0,
min=0,
max=1)
private$..crossValidate <- jmvcore::OptionBool$new(
"crossValidate",
crossValidate,
default=FALSE)
private$..nFolds <- jmvcore::OptionNumber$new(
"nFolds",
nFolds,
default=5,
min=2,
max=10)
private$..seed <- jmvcore::OptionNumber$new(
"seed",
seed,
default=12345)
self$.addOption(private$..tests)
self$.addOption(private$..testLevels)
self$.addOption(private$..gold)
self$.addOption(private$..goldPositive)
self$.addOption(private$..useCosts)
self$.addOption(private$..testCosts)
self$.addOption(private$..fpCost)
self$.addOption(private$..fnCost)
self$.addOption(private$..strategies)
self$.addOption(private$..parallelRules)
self$.addOption(private$..customThreshold)
self$.addOption(private$..maxTests)
self$.addOption(private$..sequentialStop)
self$.addOption(private$..optimizationCriteria)
self$.addOption(private$..minSensitivity)
self$.addOption(private$..minSpecificity)
self$.addOption(private$..createTree)
self$.addOption(private$..treeMethod)
self$.addOption(private$..maxDepth)
self$.addOption(private$..minSplit)
self$.addOption(private$..showAllCombinations)
self$.addOption(private$..topN)
self$.addOption(private$..compareStrategies)
self$.addOption(private$..bootstrap)
self$.addOption(private$..bootReps)
self$.addOption(private$..plotTree)
self$.addOption(private$..plotComparison)
self$.addOption(private$..plotCostEffect)
self$.addOption(private$..plotROC)
self$.addOption(private$..prevalence)
self$.addOption(private$..crossValidate)
self$.addOption(private$..nFolds)
self$.addOption(private$..seed)
}),
active = list(
tests = function() private$..tests$value,
testLevels = function() private$..testLevels$value,
gold = function() private$..gold$value,
goldPositive = function() private$..goldPositive$value,
useCosts = function() private$..useCosts$value,
testCosts = function() private$..testCosts$value,
fpCost = function() private$..fpCost$value,
fnCost = function() private$..fnCost$value,
strategies = function() private$..strategies$value,
parallelRules = function() private$..parallelRules$value,
customThreshold = function() private$..customThreshold$value,
maxTests = function() private$..maxTests$value,
sequentialStop = function() private$..sequentialStop$value,
optimizationCriteria = function() private$..optimizationCriteria$value,
minSensitivity = function() private$..minSensitivity$value,
minSpecificity = function() private$..minSpecificity$value,
createTree = function() private$..createTree$value,
treeMethod = function() private$..treeMethod$value,
maxDepth = function() private$..maxDepth$value,
minSplit = function() private$..minSplit$value,
showAllCombinations = function() private$..showAllCombinations$value,
topN = function() private$..topN$value,
compareStrategies = function() private$..compareStrategies$value,
bootstrap = function() private$..bootstrap$value,
bootReps = function() private$..bootReps$value,
plotTree = function() private$..plotTree$value,
plotComparison = function() private$..plotComparison$value,
plotCostEffect = function() private$..plotCostEffect$value,
plotROC = function() private$..plotROC$value,
prevalence = function() private$..prevalence$value,
crossValidate = function() private$..crossValidate$value,
nFolds = function() private$..nFolds$value,
seed = function() private$..seed$value),
private = list(
..tests = NA,
..testLevels = NA,
..gold = NA,
..goldPositive = NA,
..useCosts = NA,
..testCosts = NA,
..fpCost = NA,
..fnCost = NA,
..strategies = NA,
..parallelRules = NA,
..customThreshold = NA,
..maxTests = NA,
..sequentialStop = NA,
..optimizationCriteria = NA,
..minSensitivity = NA,
..minSpecificity = NA,
..createTree = NA,
..treeMethod = NA,
..maxDepth = NA,
..minSplit = NA,
..showAllCombinations = NA,
..topN = NA,
..compareStrategies = NA,
..bootstrap = NA,
..bootReps = NA,
..plotTree = NA,
..plotComparison = NA,
..plotCostEffect = NA,
..plotROC = NA,
..prevalence = NA,
..crossValidate = NA,
..nFolds = NA,
..seed = NA)
)
decisionpanelResults <- if (requireNamespace("jmvcore", quietly=TRUE)) R6::R6Class(
"decisionpanelResults",
inherit = jmvcore::Group,
active = list(
summary = function() private$.items[["summary"]],
optimalPanel = function() private$.items[["optimalPanel"]],
strategyComparison = function() private$.items[["strategyComparison"]],
individualTests = function() private$.items[["individualTests"]],
allCombinations = function() private$.items[["allCombinations"]],
treeStructure = function() private$.items[["treeStructure"]],
treePerformance = function() private$.items[["treePerformance"]],
crossValidation = function() private$.items[["crossValidation"]],
bootstrapResults = function() private$.items[["bootstrapResults"]],
treeVisualization = function() private$.items[["treeVisualization"]],
strategyComparisonPlot = function() private$.items[["strategyComparisonPlot"]],
costEffectivenessPlot = function() private$.items[["costEffectivenessPlot"]],
rocCurvesPlot = function() private$.items[["rocCurvesPlot"]],
recommendations = function() private$.items[["recommendations"]]),
private = list(),
public=list(
initialize=function(options) {
super$initialize(
options=options,
name="",
title="Decision Panel Optimization",
refs=list(
"rpart",
"partykit",
"DiagrammeR",
"ROCR",
"ClinicoPathJamoviModule"))
self$add(jmvcore::Html$new(
options=options,
name="summary",
title="Summary",
visible=TRUE))
self$add(jmvcore::Table$new(
options=options,
name="optimalPanel",
title="Optimal Test Panel",
visible=TRUE,
columns=list(
list(
`name`="rank",
`title`="Rank",
`type`="integer"),
list(
`name`="tests",
`title`="Test Combination",
`type`="text"),
list(
`name`="strategy",
`title`="Strategy",
`type`="text"),
list(
`name`="sensitivity",
`title`="Sensitivity",
`type`="number",
`format`="pc"),
list(
`name`="specificity",
`title`="Specificity",
`type`="number",
`format`="pc"),
list(
`name`="accuracy",
`title`="Accuracy",
`type`="number",
`format`="pc"),
list(
`name`="ppv",
`title`="PPV",
`type`="number",
`format`="pc"),
list(
`name`="npv",
`title`="NPV",
`type`="number",
`format`="pc"),
list(
`name`="youden",
`title`="Youden's J",
`type`="number",
`format`="zto"),
list(
`name`="cost",
`title`="Cost",
`type`="number",
`format`="zto",
`visible`="(useCosts)"),
list(
`name`="utility",
`title`="Utility",
`type`="number",
`format`="zto",
`visible`="(useCosts)"),
list(
`name`="efficiency",
`title`="Efficiency",
`type`="number",
`format`="zto",
`visible`="(useCosts)"))))
self$add(jmvcore::Table$new(
options=options,
name="strategyComparison",
title="Testing Strategy Comparison",
visible="(compareStrategies)",
columns=list(
list(
`name`="strategy",
`title`="Strategy",
`type`="text"),
list(
`name`="nTests",
`title`="Tests Used",
`type`="integer"),
list(
`name`="avgSens",
`title`="Avg Sensitivity",
`type`="number",
`format`="pc"),
list(
`name`="avgSpec",
`title`="Avg Specificity",
`type`="number",
`format`="pc"),
list(
`name`="avgAccuracy",
`title`="Avg Accuracy",
`type`="number",
`format`="pc"),
list(
`name`="avgCost",
`title`="Avg Cost",
`type`="number",
`format`="zto",
`visible`="(useCosts)"),
list(
`name`="bestPanel",
`title`="Best Panel",
`type`="text"),
list(
`name`="bestPerformance",
`title`="Best Performance",
`type`="number",
`format`="zto"))))
self$add(jmvcore::Table$new(
options=options,
name="individualTests",
title="Individual Test Performance",
visible=TRUE,
columns=list(
list(
`name`="test",
`title`="Test",
`type`="text"),
list(
`name`="sensitivity",
`title`="Sensitivity",
`type`="number",
`format`="pc"),
list(
`name`="specificity",
`title`="Specificity",
`type`="number",
`format`="pc"),
list(
`name`="accuracy",
`title`="Accuracy",
`type`="number",
`format`="pc"),
list(
`name`="ppv",
`title`="PPV",
`type`="number",
`format`="pc"),
list(
`name`="npv",
`title`="NPV",
`type`="number",
`format`="pc"),
list(
`name`="plr",
`title`="+LR",
`type`="number",
`format`="zto"),
list(
`name`="nlr",
`title`="-LR",
`type`="number",
`format`="zto"),
list(
`name`="cost",
`title`="Cost",
`type`="number",
`format`="zto",
`visible`="(useCosts)"))))
self$add(jmvcore::Table$new(
options=options,
name="allCombinations",
title="All Test Combinations",
visible="(showAllCombinations)",
columns=list(
list(
`name`="tests",
`title`="Tests",
`type`="text"),
list(
`name`="strategy",
`title`="Strategy",
`type`="text"),
list(
`name`="nTests",
`title`="N Tests",
`type`="integer"),
list(
`name`="sensitivity",
`title`="Sensitivity",
`type`="number",
`format`="pc"),
list(
`name`="specificity",
`title`="Specificity",
`type`="number",
`format`="pc"),
list(
`name`="accuracy",
`title`="Accuracy",
`type`="number",
`format`="pc"),
list(
`name`="ppv",
`title`="PPV",
`type`="number",
`format`="pc"),
list(
`name`="npv",
`title`="NPV",
`type`="number",
`format`="pc"),
list(
`name`="cost",
`title`="Cost",
`type`="number",
`format`="zto",
`visible`="(useCosts)"),
list(
`name`="utility",
`title`="Utility",
`type`="number",
`format`="zto",
`visible`="(useCosts)"))))
self$add(jmvcore::Html$new(
options=options,
name="treeStructure",
title="Decision Tree Structure",
visible="(createTree)"))
self$add(jmvcore::Table$new(
options=options,
name="treePerformance",
title="Decision Tree Performance",
visible="(createTree)",
columns=list(
list(
`name`="node",
`title`="Node",
`type`="text"),
list(
`name`="condition",
`title`="Condition",
`type`="text"),
list(
`name`="nCases",
`title`="N Cases",
`type`="integer"),
list(
`name`="probPositive",
`title`="P(Disease+)",
`type`="number",
`format`="pc"),
list(
`name`="decision",
`title`="Decision",
`type`="text"),
list(
`name`="accuracy",
`title`="Node Accuracy",
`type`="number",
`format`="pc"))))
self$add(jmvcore::Table$new(
options=options,
name="crossValidation",
title="Cross-Validation Results",
visible="(crossValidate)",
columns=list(
list(
`name`="panel",
`title`="Test Panel",
`type`="text"),
list(
`name`="strategy",
`title`="Strategy",
`type`="text"),
list(
`name`="cvSensitivity",
`title`="CV Sensitivity",
`type`="number",
`format`="pc"),
list(
`name`="cvSensitivitySD",
`title`="SD",
`type`="number",
`format`="zto"),
list(
`name`="cvSpecificity",
`title`="CV Specificity",
`type`="number",
`format`="pc"),
list(
`name`="cvSpecificitySD",
`title`="SD",
`type`="number",
`format`="zto"),
list(
`name`="cvAccuracy",
`title`="CV Accuracy",
`type`="number",
`format`="pc"),
list(
`name`="cvAccuracySD",
`title`="SD",
`type`="number",
`format`="zto"))))
self$add(jmvcore::Table$new(
options=options,
name="bootstrapResults",
title="Bootstrap Confidence Intervals",
visible="(bootstrap)",
columns=list(
list(
`name`="panel",
`title`="Test Panel",
`type`="text"),
list(
`name`="metric",
`title`="Metric",
`type`="text"),
list(
`name`="estimate",
`title`="Estimate",
`type`="number",
`format`="zto"),
list(
`name`="lower",
`title`="2.5%",
`type`="number",
`format`="zto"),
list(
`name`="upper",
`title`="97.5%",
`type`="number",
`format`="zto"),
list(
`name`="se",
`title`="SE",
`type`="number",
`format`="zto"))))
self$add(jmvcore::Image$new(
options=options,
name="treeVisualization",
title="Decision Tree",
width=800,
height=600,
renderFun=".plotTree",
visible="(plotTree && createTree)",
clearWith=list(
"tests",
"gold",
"goldPositive",
"treeMethod",
"maxDepth",
"minSplit")))
self$add(jmvcore::Image$new(
options=options,
name="strategyComparisonPlot",
title="Strategy Comparison",
width=700,
height=500,
renderFun=".plotStrategyComparison",
visible="(plotComparison && compareStrategies)",
clearWith=list(
"tests",
"strategies",
"optimizationCriteria")))
self$add(jmvcore::Image$new(
options=options,
name="costEffectivenessPlot",
title="Cost-Effectiveness Frontier",
width=700,
height=500,
renderFun=".plotCostEffectiveness",
visible="(plotCostEffect && useCosts)",
clearWith=list(
"tests",
"testCosts",
"fpCost",
"fnCost")))
self$add(jmvcore::Image$new(
options=options,
name="rocCurvesPlot",
title="ROC Curves for Top Panels",
width=700,
height=500,
renderFun=".plotROCCurves",
visible="(plotROC)",
clearWith=list(
"tests",
"topN")))
self$add(jmvcore::Html$new(
options=options,
name="recommendations",
title="Recommendations",
visible=TRUE))}))
decisionpanelBase <- if (requireNamespace("jmvcore", quietly=TRUE)) R6::R6Class(
"decisionpanelBase",
inherit = jmvcore::Analysis,
public = list(
initialize = function(options, data=NULL, datasetId="", analysisId="", revision=0) {
super$initialize(
package = "ClinicoPath",
name = "decisionpanel",
version = c(0,0,3),
options = options,
results = decisionpanelResults$new(options=options),
data = data,
datasetId = datasetId,
analysisId = analysisId,
revision = revision,
pause = NULL,
completeWhenFilled = FALSE,
requiresMissings = FALSE,
weightsSupport = 'auto')
}))
#' Decision Panel Optimization
#'
#' Optimize diagnostic test panels by evaluating various combination
#' strategies including parallel testing (cotest), sequential testing, and
#' repeated tests. Creates decision trees to minimize cost while maximizing
#' accuracy.
#'
#'
#' @examples
#' \donttest{
#' # example will be added
#'}
#' @param data The data as a data frame.
#' @param tests Variables representing different diagnostic tests. Can select
#' multiple tests (up to 10) for panel optimization.
#' @param testLevels String specifying positive test levels. Can be: - Single
#' value (used for all tests): "Positive" - Multiple values (comma-separated):
#' "Positive,Yes,1" - Empty string (auto-detection): ""
#' @param gold The gold standard variable for disease classification.
#' @param goldPositive The level of the gold standard variable that indicates
#' disease presence.
#' @param useCosts Include cost considerations in the optimization process.
#' @param testCosts Comma-separated costs for each test in the same order as
#' selected tests. Example: "10,25,50" for three tests.
#' @param fpCost Cost or harm associated with a false positive result.
#' @param fnCost Cost or harm associated with a false negative result.
#' @param strategies Which testing strategies to evaluate in the analysis.
#' @param parallelRules Rule for combining results in parallel testing.
#' @param customThreshold Number of positive tests required for overall
#' positive result (when using custom rule).
#' @param maxTests Maximum number of tests to combine in any strategy.
#' @param sequentialStop When to stop testing in sequential strategies.
#' @param optimizationCriteria Primary criterion for optimizing test panels.
#' @param minSensitivity Minimum sensitivity constraint for panel selection.
#' @param minSpecificity Minimum specificity constraint for panel selection.
#' @param createTree Generate an optimal decision tree for test sequencing.
#' @param treeMethod Method for constructing the decision tree.
#' @param maxDepth Maximum depth of the decision tree.
#' @param minSplit Minimum number of observations required to split a node.
#' @param showAllCombinations Display performance metrics for all possible
#' test combinations.
#' @param topN Number of best-performing panels to display in results.
#' @param compareStrategies Show comparative analysis of different testing
#' strategies.
#' @param bootstrap Calculate bootstrap confidence intervals for performance
#' metrics.
#' @param bootReps Number of bootstrap replications for confidence intervals.
#' @param plotTree Display visual representation of the optimal decision tree.
#' @param plotComparison Create comparison plots for different testing
#' strategies.
#' @param plotCostEffect Create cost-effectiveness frontier plot.
#' @param plotROC Display ROC curves for top performing panels.
#' @param prevalence Known disease prevalence (0 = use sample prevalence).
#' @param crossValidate Perform k-fold cross-validation for panel performance.
#' @param nFolds Number of folds for cross-validation.
#' @param seed Random seed for reproducibility in bootstrap and
#' cross-validation.
#' @return A results object containing:
#' \tabular{llllll}{
#' \code{results$summary} \tab \tab \tab \tab \tab a html \cr
#' \code{results$optimalPanel} \tab \tab \tab \tab \tab a table \cr
#' \code{results$strategyComparison} \tab \tab \tab \tab \tab a table \cr
#' \code{results$individualTests} \tab \tab \tab \tab \tab a table \cr
#' \code{results$allCombinations} \tab \tab \tab \tab \tab a table \cr
#' \code{results$treeStructure} \tab \tab \tab \tab \tab a html \cr
#' \code{results$treePerformance} \tab \tab \tab \tab \tab a table \cr
#' \code{results$crossValidation} \tab \tab \tab \tab \tab a table \cr
#' \code{results$bootstrapResults} \tab \tab \tab \tab \tab a table \cr
#' \code{results$treeVisualization} \tab \tab \tab \tab \tab an image \cr
#' \code{results$strategyComparisonPlot} \tab \tab \tab \tab \tab an image \cr
#' \code{results$costEffectivenessPlot} \tab \tab \tab \tab \tab an image \cr
#' \code{results$rocCurvesPlot} \tab \tab \tab \tab \tab an image \cr
#' \code{results$recommendations} \tab \tab \tab \tab \tab a html \cr
#' }
#'
#' Tables can be converted to data frames with \code{asDF} or \code{\link{as.data.frame}}. For example:
#'
#' \code{results$optimalPanel$asDF}
#'
#' \code{as.data.frame(results$optimalPanel)}
#'
#' @export
decisionpanel <- function(
data,
tests,
testLevels = "",
gold,
goldPositive,
useCosts = FALSE,
testCosts = "",
fpCost = 100,
fnCost = 1000,
strategies = "all",
parallelRules = "any",
customThreshold = 2,
maxTests = 3,
sequentialStop = "positive",
optimizationCriteria = "accuracy",
minSensitivity = 0.8,
minSpecificity = 0.8,
createTree = FALSE,
treeMethod = "costSensitive",
maxDepth = 5,
minSplit = 20,
showAllCombinations = FALSE,
topN = 10,
compareStrategies = TRUE,
bootstrap = FALSE,
bootReps = 1000,
plotTree = TRUE,
plotComparison = TRUE,
plotCostEffect = TRUE,
plotROC = FALSE,
prevalence = 0,
crossValidate = FALSE,
nFolds = 5,
seed = 12345) {
if ( ! requireNamespace("jmvcore", quietly=TRUE))
stop("decisionpanel requires jmvcore to be installed (restart may be required)")
if ( ! missing(tests)) tests <- jmvcore::resolveQuo(jmvcore::enquo(tests))
if ( ! missing(gold)) gold <- jmvcore::resolveQuo(jmvcore::enquo(gold))
if (missing(data))
data <- jmvcore::marshalData(
parent.frame(),
`if`( ! missing(tests), tests, NULL),
`if`( ! missing(gold), gold, NULL))
for (v in tests) if (v %in% names(data)) data[[v]] <- as.factor(data[[v]])
for (v in gold) if (v %in% names(data)) data[[v]] <- as.factor(data[[v]])
options <- decisionpanelOptions$new(
tests = tests,
testLevels = testLevels,
gold = gold,
goldPositive = goldPositive,
useCosts = useCosts,
testCosts = testCosts,
fpCost = fpCost,
fnCost = fnCost,
strategies = strategies,
parallelRules = parallelRules,
customThreshold = customThreshold,
maxTests = maxTests,
sequentialStop = sequentialStop,
optimizationCriteria = optimizationCriteria,
minSensitivity = minSensitivity,
minSpecificity = minSpecificity,
createTree = createTree,
treeMethod = treeMethod,
maxDepth = maxDepth,
minSplit = minSplit,
showAllCombinations = showAllCombinations,
topN = topN,
compareStrategies = compareStrategies,
bootstrap = bootstrap,
bootReps = bootReps,
plotTree = plotTree,
plotComparison = plotComparison,
plotCostEffect = plotCostEffect,
plotROC = plotROC,
prevalence = prevalence,
crossValidate = crossValidate,
nFolds = nFolds,
seed = seed)
analysis <- decisionpanelClass$new(
options = options,
data = data)
analysis$run()
analysis$results
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.