These data are used in the book "Computational Statistics" 
by G.H. Givens and J.A. Hoeting.

Name: face-recognition.dat
Chapter: 2, 12
Discussed in: Example 2.5, Problem 12.4

Source:  Geof Givens and Ross Beveridge, Colorado State University

Description: These data arise from the evaluation of a human face
recognition algorithm. Details of the experiment are given in the
text.  The variables relate to pairs of images of the same person, one
of which is the probe image.  The response variable is "match", with
match=1 indicating that the probe image was correctly matched to the
other image of the same person and match=0 indicating that the probe
was incorrectly matched to an image of someone else.  These predictors
include a measure of pixel intensity differences in the eye region
(eyediff) and in the nose/cheek region (nosecheekdiff), and a measure
of comparative pixel intensity variability between the two images of
the same person (variabilityratio). 

** Note: these data are slightly different than the dataset used in
the book example, but the result should be very close.

Variable labels: match, eyediff, nosecheekdiff, variabilityratio