hw: Articial data with classical outliers.
In p3d: Plot 3d data and fitted surfaces

Description Usage Format Details Examples

Articial data on health as predicted by height and weight with outliers to show the effect of different types of outliers.

1
2
3

data(hw)

data(hwoutliers)

A data frame with 4 datasets, indexed by Type, each with 15 'good' observations. The last 3 datasets eadh contain, in addition to the 'good' data, an outlier.

Weight: a measure of body weight
Height: a measure of height
Health: an index of health: higher is better
Type: numerical type, 0, 1, 2 or 3
Outlier: a factor with levels none Type 1 Type 2 Type 3

hw has four datasets identified by Outlier indicating the 'type' of outlier depicted in each dataset: none, Type 1: large residual for y with typical value of x, Type 2: small residual for y with atypical value of x, and Type 3: atypical x and poorly fitting y. hwoutliers contains the rows of 'acceptable' data and one row for each of the three types of outlier. This format lends itself better to 3-D plots. Note that this nomenclature for outlier is not widely adopted!