designTreatmentsC: designTreatmentsC

Description

Function to design variable treatments for binary prediction of a categorical outcome. Data frame is assumed to have only atomic columns except for dates (which are converted to numeric).

Usage

designTreatmentsC(dframe, varlist, outcomename, outcometarget, weights = c(),
  minFraction = 0.02, smFactor = 0, rareCount = 2, rareSig = 0.3,
  maxMissing = 0.04, collarProb = 0, returnXFrame = FALSE,
  scale = FALSE, doCollar = TRUE, verbose = TRUE,
  parallelCluster = NULL)

Arguments

dframe

Data frame to learn treatments from (training data).

varlist

Names of columns to treat (effective variables).

outcomename

Name of column holding outcome variable.

outcometarget

Value/level of outcome to be considered "success"

weights

optional training weights for each row

minFraction

optional minimum frequency a categorical level must have to be converted to an indicator column.

smFactor

optional smoothing factor for impact coding models.

rareCount

optional integer, supress direct effects of level of this count or less.

rareSig

optional integer, supress direct effects of level of this signficance or less.

maxMissing

optional maximum fraction (by data weight) of a categorical variable that are allowed before switching from indicators to impact coding.

collarProb

what fraction of the data (pseudo-probability) to collar data at (

returnXFrame

optional if TRUE return out of sample transformed frame.

scale

logical optional controls scaling for scoring and returnXFrame

doCollar

logical optional controls collaring for scoring and returnXFrame

verbose

if TRUE print progress.

parallelCluster

(optional) a cluster object created by package parallel or package snow

Value

treatment plan (for use with prepare)

Details

The main fields are mostly vectors with names (all with the same names in the same order):

- vars : (character array without names) names of variables (in same order as names on the other diagnostic vectors) - varMoves : logical TRUE if the variable varied during hold out scoring, only variables that move will be in the treated frame - #' - sig : an estimate signficance of effect

See the vtreat vignette for a bit more detail and a worked example.

Examples

Run this code

dTrainC <- data.frame(x=c('a','a','a','b','b','b'),
   z=c(1,2,3,4,5,6),
   y=c(FALSE,FALSE,TRUE,FALSE,TRUE,TRUE))
dTestC <- data.frame(x=c('a','b','c',NA),
   z=c(10,20,30,NA))
treatmentsC <- designTreatmentsC(dTrainC,colnames(dTrainC),'y',TRUE)
dTrainCTreated <- prepare(treatmentsC,dTrainC,pruneSig=0.99)
dTestCTreated <- prepare(treatmentsC,dTestC,pruneSig=0.99)