Record Levels of Factors Used in a Model

.getXlevels

Description

Records the levels of the factors used as predictors in a model.

Usage

.getXlevels(Terms, m)

Arguments

Terms	an object of class terms as made by the terms function.
m	a data.frame containing the variables used in the Terms argument.

Details

This function is used in modeling functions (for example, lm) where it is typically used to create the xlevels component of the return value. The predict method for the modeling function should use this component to map character strings to integers when converting the newdata to a data.frame similar to the data argument used to fit the model. (predict.lm does this by calling model.matrix.lm.)

Many modeling algorithms use character predictors directly without mapping them to an integer, so they have no need for .getXlevels.

Users of modeling functions have no need of .getXlevels; rather, it is for writers of modeling functions.

Value

returns a list with one component for each character-valued or factor-valued predictor listing the levels in that predictor.

See Also

model.matrix.default (the argument is supplied only when called from model.matrix.lm

Examples

myFormula <- y ~ x1 + x2
myDataFrame <- data.frame(
    y=1:10,
    x1=log(1:10),
    x2=factor(rep(c("B","C"),c(4,6)), levels=c("C","B","A")),
    x3=factor(rep(c("P","Q","R"),c(3,3,4))))
.getXlevels(terms(myFormula),
    model.frame(myFormula, drop.unused.levels = TRUE, data = myDataFrame))

Package stats version 6.1.1-7
Package Index