ppwdeming - Precision Profile Weighted Deming Regression
Weighted Deming regression, also known as
'errors-in-variable' regression, is applied with suitable
weights. Weights are modeled via a precision profile; thus the
methods implemented here are referred to as precision profile
weighted Deming (PWD) regression. The package covers two
settings – one where the precision profiles are known either
from external studies or from adequate replication of the X and
Y readings, and one in which there is a plausible functional
form for the precision profiles but the exact (unknown)
function must be estimated from the (generally singlicate)
readings. The function set includes tools for: estimated
standard errors (via jackknifing); standardized-residual
analysis function with regression diagnostic tools for
normality, linearity and constant variance; and an outlier
analysis identifying significant outliers for closer
investigation. The following reference provides further
information on mathematical derivations and applications.
Hawkins, D.M., and J.J. Kraker (2026). 'Precision Profile
Weighted Deming Regression for Methods Comparison'. The Journal
of Applied Laboratory Medicine 11, 379-392
<doi:10.1093/jalm/jfaf183>. Weighted Deming regression is also
now implemented for multiple instruments , as set out in
Hawkins, D.M., and J.J. Kraker (2026). 'Multiple Instrument
Methods Comparison by Precision weighted Deming Regression', on
Arxiv <doi:10.48550/arXiv.2607.11776>. The “multi” functions
refer to the multiple instrument analysis, and the "PWD"
functions refer to the two-instrument.
