ModelEnv-class: Class "ModelEnv"
In modeltools: Tools and Classes for Statistical Models

Description Details Objects from the Class Slots Methods Examples

A class for model environments.

Objects of class ModelEnv basically consist of an environment for data storage as well as get and set methods.

na.fail returns FALSE when at least one missing value occurs in object@env. na.pass returns object unchanged and na.omit returns a copy of object with all missing values removed.

Objects can be created by calls of the form new("ModelEnv", ...).

env:: Object of class "environment".
get:: Object of class "function" for extracting objects from environment env.
set:: Object of class "function" for setting object in environment env.
hooks:: A list of hook collections.

clone: signature(object = "ModelEnv"): copy an object.
dimension: signature(object = "ModelEnv", which = "character"): get the dimension of an object.
empty: signature(object = "ModelEnv"): Return TRUE, if the model environment contains no data.
has: signature(object = "ModelEnv", which = "character"): check if an object which is available in env.
initialize: signature(.Object = "ModelEnv"): setup new objects.
show: signature(object = "ModelEnv"): show object.
subset: signature(x = "ModelEnv"): extract subsets from an object.
na.pass: na.action method for ModelEnv objects.
na.fail: na.action method for ModelEnv objects.
na.omit: na.action method for ModelEnv objects.

### a new object
me <- new("ModelEnv")

## the new model environment is empty
empty(me)

### define a bivariate response variable
me@set("response", data.frame(y = rnorm(10), x = runif(10)))
me

## now it is no longer empty
empty(me)

### check if a response is available
has(me, "response")

### the dimensions
dimension(me, "response")

### extract the data
me@get("response")

df <- data.frame(x = rnorm(10), y = rnorm(10))

## hook for set method:
mf <- ModelEnvFormula(y ~ x-1, data = df, setHook=list(designMatrix=scale))
mf@get("designMatrix")
mf@set(data=df[1:5,])
mf@get("designMatrix")

### NA handling
df$x[1] <- NA
mf <- ModelEnvFormula(y ~ x, data = df, na.action = na.pass)
mf
na.omit(mf)