FunctionXform: Add a function transformation to a WrapData object.
In pmmlTransformations: Transforms Input Data from a PMML Perspective

Description Usage Arguments Details Value Author(s) See Also Examples

Add a function transformation to a WrapData object.

1 2	FunctionXform(boxdata, origFieldName, newFieldName = "newField", newFieldDataType = "numeric", formulaText, mapMissingTo = NA)

`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

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.

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.

Dmitriy Bolotov

WrapData

# 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)