Clinical Classification

Usage

classification(
  data,
  dep,
  indep,
  testSize = 0.33,
  noOfFolds = 10,
  testing,
  reporting = list("classifMetrices"),
  classifier,
  minSplit = 20,
  minBucket = 0,
  complexity = 0.01,
  maxCompete = 4,
  maxSurrogate = 5,
  unsurrogate = 2,
  noCrossValidations = 10,
  maxDepth = 30,
  noOfTrees = 10,
  maxDepthRandFor = 30,
  sampleFraction = 1,
  splitRule,
  knnNeighbors = 5,
  knnDistance = "euclidean",
  svmKernel = "radial",
  svmCost = 1,
  svmGamma = 1,
  plotDecisionTree = FALSE,
  predictedFreq = FALSE,
  printRandForest = FALSE,
  predictedFreqRF = FALSE,
  balancingMethod = "none",
  clinicalCutoff = 0.5,
  validateMethod = "holdout",
  bootstrapSamples = 1000,
  reportClinicalMetrics = TRUE,
  reportConfidenceIntervals = TRUE
)

Arguments

data: .
dep: .
indep: .
testSize: .
noOfFolds: .
testing: .
reporting: .
classifier: .
minSplit: .
minBucket: .
complexity: .
maxCompete: .
maxSurrogate: .
unsurrogate: .
noCrossValidations: .
maxDepth: .
noOfTrees: .
maxDepthRandFor: .
sampleFraction: .
splitRule: .
knnNeighbors: Number of nearest neighbors for KNN classification.
knnDistance: .
svmKernel: .
svmCost: Regularization parameter for SVM.
svmGamma: Kernel coefficient for SVM (used in RBF, polynomial, sigmoid kernels).
plotDecisionTree: .
predictedFreq: .
printRandForest: .
predictedFreqRF: .
balancingMethod: Method for handling class imbalance in medical datasets.
clinicalCutoff: Probability threshold for clinical decision making.
validateMethod: Validation method for clinical model assessment.
bootstrapSamples: Number of bootstrap samples for confidence intervals.
reportClinicalMetrics: Report sensitivity, specificity, PPV, NPV, and likelihood ratios.
reportConfidenceIntervals: Include 95\ clinical metrics.

Value

A results object containing:

`results$modelSettings`					a html
`results$confusion$matrix`					a table
`results$classificationMetrics$general`					a table
`results$classificationMetrics$clinicalMetrics`					a table
`results$classificationMetrics$class`					a table
`results$rocCurvePlot`					an image
`results$decisionTreeModel`					an image
`results$predictedFreqPlot`					an image
`results$printRandForest$randomForestModel`					a table
`results$text`					a preformatted