FunctionXform {pmmlTransformations}R Documentation

Add a function transformation to a WrapData object.

Description

Add a function transformation to a WrapData object.

Usage

FunctionXform(boxdata, origFieldName, newFieldName = "newField",
  newFieldDataType = "numeric", formulaText, mapMissingTo = NA)

Arguments

boxdata

wrapper object obtained by using the WrapData function on raw data

origFieldName

string specifying name(s) of the original data field(s) being used in the transformation

newFieldName

name of the new field created by the transformation

newFieldDataType

data type of the new field created by the transformation

formulaText

string expression specifying the transformation

mapMissingTo

value to be given to the transformed variable if the value of any input variable is missing

Details

Calculate the expression provided in formulaText for every row in the boxdata$data data frame. The formulaText argument must represent a valid R expression, and any functions used in formulaText must be defined in the current environment.

The name of the new field is optional (a default name is provided), but an error will be thrown if attempting to create a field with a name that already exists in the WrapData object.

Value

R object containing the raw data, the transformed data and data statistics. The data data frame will contain a new newFieldName column, and fieldData will contain a new newFieldName row.

Author(s)

Dmitriy Bolotov

See Also

WrapData

Examples

# Load the standard iris dataset
data(iris)

# Wrap the data
irisBox <- WrapData(iris)

# Perform a transform on the Sepal.Length field: 
# the value is squared and then divided by 100
irisBox <- FunctionXform(irisBox,origFieldName="Sepal.Length",
                         newFieldName="Sepal.Length.Transformed",
                         formulaText="(Sepal.Length^2)/100")

# Combine two fields to create another new feature:                      
irisBox <- FunctionXform(irisBox,
                         origFieldName="Sepal.Width, Petal.Width",
                         newFieldName="Width.Sum",
                         formulaText="Sepal.Width + Sepal.Length")
                         
# Create linear model using the derived features
fit <- lm(Petal.Length ~ 
         Sepal.Length.Transformed + Width.Sum, data=irisBox$data)

# Create pmml from the fit
# library(pmml)
# fit_pmml <- pmml(fit, transform=irisBox)

[Package pmmlTransformations version 1.3.3 Index]