bse-class: An S4 class for broken stick estimates

In stefvanbuuren/minihealth: Mini Dossier for Individual Health Data

Description

The bse class stores three variables that result from predicting a given broken stick model. The bse class extends the xyz-class class. The bse class calculates predicted values from the broken stick model fitted to the values specified in the data argument of the new("bse", data) function.

Slots

call:

An object of class call that refers to an object with a pre-fitted broken stick model. Evaluation by eval(call) should yield an object of class brokenstick as created by brokenstick.

found:

Logical indicating whether the broken stick model specified by call was actually found.

zscale:

A logical indicating whether the broken stick model models the outcome in the Z-scale (zscale = TRUE, default) or Y-scale (zscale = FALSE). At initialization, the new() function tries to infer and set this flag from the model specified by call.

Author(s)

Stef van Buuren

See Also

xyz-class, brokenstick, predict.brokenstick

Examples

library(donorloader)
smocc_bs <- load_data(dnr = "smocc_bs")

# specify three height measures for child
child <- new("xyz", x = c(0, 0.2, 0.5), y = c(51.0, 54.1, 63.4))
# we get no Z-score conversion because haven't told the child's sex
child

boy <- new("xyz", x = c(0, 0.2, 0.5), y = c(51.0, 54.1, 63.4), sex = "male")
girl <- new("xyz", x = c(0, 0.2, 0.5), y = c(51.0, 54.1, 63.4), sex = "female")
boy
girl

d1 <- new("bse", child)
d1

# Same, but now for the girl
d2 <- new("bse", girl)
d2

# Calculate predicted value for each x
d3 <- new("bse", new("xyz", x = c(0, 0.2, 0.5), y = c(51.0, 54.1, 63.4),
sex = "female"))
d3

d4 <- new("bse", new("xyz", x = c(0, 0.2, 0.5), y = c(51.0, 54.1, 63.4),
sex = "female"), at = "knots")
d4

# use transform_y function instead of reference call
# now, -0.126 is converted to 41.2 cm
d5 <- new("bse", girl, usetransform = TRUE, ga = 32, sex = "female")
d5
d5@transform

stefvanbuuren/minihealth documentation built on March 11, 2021, 7:10 p.m.