multresp: Create an indicator matrix from a multiple response variable

In gmonette/spida15: Collection of miscellaneous functions for mixed models etc. prepared for SPIDA 2009+ (development version)

Description

A multiple reponse variable is one whose values consist of a list of possible choices. Such a variable can be represented in a linear model through a matrix of indicator variables.

Usage

multresp(x, split = "[ \t,]+", sep = ".", root = deparse(substitute(x)))

Arguments

x	a factor or character vector containing comma- or blank-separated lists of reponse choices for each subject. For example, each subject could have a list of symptoms: c('cough fever', 'cough bronchitis', 'headache wheeze', ”, 'cough high_fever', NA).
split	a regular expression matching the separator between fields. The default is "[ \t,]+" which matches one more blank, tab or comma. To use semicolons instead of commas as separators, use: "[ \t;]+" .
sep	the separator used to generated the column names of the indicator matrix. sep separates the root from the response values. Default: ".". For example, if x is the name of the variable, then the default column name for response value 'cough' is x.cough.
root	the root for column names. By default, the name of the variable. the

Value

A matrix of indicator variables. For, example, if x <-c('cough fever', 'cough bronchitis', 'headache wheeze', '', 'cough high_fever', NA) multresp(x) produces: x.bronchitis x.cough x.fever x.headache x.high_fever x.wheeze [1,] 0 1 1 0 0 0 [2,] 1 1 0 0 0 0 [3,] 0 0 0 1 0 1 [4,] 0 0 0 0 0 0 [5,] 0 1 0 0 1 0 [6,] 0 0 0 0 0 0

Author(s)

G. Monette ([e-mail])

Examples

x <-c('cough fever', 'cough bronchitis', 'headache wheeze', '', 'cough high_fever', NA)
multresp(x)

data <- data.frame(x = x, y = 1:6)
data <- with(data, cbind(data, multresp(x)))
data

gmonette/spida15 documentation built on May 17, 2019, 7:26 a.m.