Diagnostics for Conformity of Paired Quantitative Measurements
(Hawkins, D. M., Statistics in Medicine, 21, 2002, 1913--1935.)
The data sets used in this paper can be downloaded as an Excel spreadsheet by clicking here.
Some Issues in Resolution of Diagnostic Tests using an Imperfect Gold Standard
Hawkins, D. M.,
Garrett, J. A. and Stephenson, B., (2001), Statistics
in Medicine, 20, 1987-2001.
A version of the paper correcting several typographic errors
can be downloaded here.
Inconsistency of Resampling Algorithms for High Breakdown Regression Estimators and a New Algorithm (with discussion)
(Hawkins, D. M. and Olive, D, J. Journal of the American Statistical Association. 97, 2002, 136 - 148, rejoinder 156-159.)
The data sets used
in this paper can be downloaded as CSV files by clicking here.
the
Gladstone data;
the Buxton data.
Three papers dealing with Phase II uses of a change point formulation have
appeared. All refer to extensive tables of control limits for the relevant
chart statistic; these tables are provided below.
All use the same format: you ignore 'skip' observations, and start testing with
the following observation. The tables list for different values of 'skip' (in
the first column) the control limits for each subsequent observation (in the
second column.)
The 'data' columns of the tables correspond to false alarm rates listed in the
first header line. For example, the column headed p002 has a false alarm
probability of 0.002 at each observation, and so has an in-control ARL of 500.
The publications, and the tables of control
limits, are:
'The Changepoint Model for Statistical Process
Control', Hawkins, D. M., Qiu, P., and Kang, C.-W.
2003, Journal of Quality Technology, 35, 355-365.
These limits are newly computed using 40 million sequences, followed by some
smoothing. The listed entries are all believed to be accurate to a maximum
error of 3 in the third decimal.
control
limits
'Change Point Model for Statistical Process
Control With Shift in Variance', Hawkins, D. M., and Zamba, K. D., 2005, Journal of Quality Technology, 37,
21-31.
The table is as outlined in the publication. The second line of each pair gives
estimates of the standard errors of the control limits.
control
limits
'Statistical Process Control for Shift in Mean or Variance Using the Change
Point Formulation', Hawkins, D. M., and Zamba, K. D.,
2004, Technometrics, 47, 164-173.
control
limits
Zamba, K. D., and Hawkins, D. M., 'A Multivariate
Change-Point Model for Statistical Processs Control',
2006, Technometrics. 48, 539-549
control limits
This link connects to the Fortran 95 source program and Windows executable for the methodology described in TR 641.
This link is a PC executable for
calculating the control limit for a MEWMC chart.
Fortran 95 source code for the executable can be used
for other hardware.
Code for
simulating CUSUM and EWMA.