merror3.0 package

Accuracy and Precision of Measurements

alpha.beta.sigma

Build an alpha-beta-sigma Matrix for Use with the cplot Function

beta.bar

Compute the estimates of betas.

cb.pd

Compute accuracy estimates and maximum likelihood estimates of precisi...

cplot

Scatter plot of observations for a pair of devices with calibration cu...

errors.cb

Extracts the estimated measurement errors assuming there is a constant...

errors.nb

Extracts the estimated measurement errors assuming there is no bias an...

errors.ncb

Extracts the estimated measurement errors assuming there is a nonconst...

lrt

Likelihood ratio test for all betas equalling one.

merror.pairs

A modified "pairs" plot with all axes haveing the same range.

mle

Compute maximum likelihood estimates of precision.

mle.se2

Compute squared standard errors for imprecision estimates for the cons...

ncb.od

Compute accuracy estimates and maximum likelihood estimates of precisi...

omx

Compute full information maximum likelihood estimates of accuracy and ...

panel.merror

Draw diagonal line (line of equality) on merror.pairs plots

precision.grubbs.cb.pd

Computes Grubbs' method of moments estimators of precision for the con...

precision.grubbs.ncb.od

Computes Grubbs' method of moments estimators of precision for the non...

precision.mle.ncb.od

Computes iterative approximation to mle precision estimates for noncon...

process.sd

Compute process standard deviation

process.var.mle.jaech.err

Compute process variance but with minor error in Jaech Fortran code.

process.var.mle

Compute process variance.

sigma_mle

Computes the ith iteration for computing the squared imprecision estim...

N>=3 methods are used to measure each of n items. The data are used to estimate simultaneously systematic error (bias) and random error (imprecision). Observed measurements for each method or device are assumed to be linear functions of the unknown true values and the errors are assumed normally distributed. Pairwise calibration curves and plots can be easily generated. Unlike the 'ncb.od' function, the 'omx' function builds a one-factor measurement error model using 'OpenMx' and allows missing values, uses full information maximum likelihood to estimate parameters, and provides both likelihood-based and bootstrapped confidence intervals for all parameters, in addition to Wald-type intervals.

  • Maintainer: Richard A. Bilonick
  • License: GPL (>= 2)
  • Last published: 2023-08-29