diff options
Diffstat (limited to 'man')
| -rw-r--r-- | man/mkinfit.Rd | 9 | ||||
| -rw-r--r-- | man/sigma_twocomp.Rd (renamed from man/sigma_rl.Rd) | 18 | ||||
| -rw-r--r-- | man/synthetic_data_for_UBA.Rd | 5 |
3 files changed, 23 insertions, 9 deletions
diff --git a/man/mkinfit.Rd b/man/mkinfit.Rd index 32edb28b..0f06b321 100644 --- a/man/mkinfit.Rd +++ b/man/mkinfit.Rd @@ -212,11 +212,14 @@ mkinfit(mkinmod, observed, \code{reweight.tol} or up to the maximum number of iterations specified by \code{reweight.max.iter}. The second reweighting method is called "tc" (two-component error model). - When using this method, the two components of the error model according - to Rocke and Lorenzato (1995) are estimated from the fit and the resulting + When using this method, the two components an error model similar to + Rocke and Lorenzato (1995) are estimated from the fit and the resulting variances are used for weighting the residuals in each iteration until convergence of these components or up to the maximum number of iterations - specified. + specified. Note that this method deviates from the model by Rocke and + Lorenzato, as their model implies that the errors follow a lognormal + distribution for large values, not a normal distribution as assumed by this + method. } \item{reweight.tol}{ Tolerance for convergence criterion for the variance components diff --git a/man/sigma_rl.Rd b/man/sigma_twocomp.Rd index 0b5d6f3c..6f941093 100644 --- a/man/sigma_rl.Rd +++ b/man/sigma_twocomp.Rd @@ -1,15 +1,19 @@ -\name{sigma_rl} -\alias{sigma_rl} -\title{ Two component error model of Rocke and Lorenzato} +\name{sigma_twocomp} +\alias{sigma_twocomp} +\title{Two component error model} \description{ Function describing the standard deviation of the measurement error in dependence of the measured value \eqn{y}: \deqn{\sigma = \sqrt{ \sigma_{low}^2 + y^2 * {rsd}_{high}^2}}{% sigma = sqrt(sigma_low^2 + y^2 * rsd_high^2)} + + This is the error model used for example by Werner et al. (1978). The model + proposed by Rocke and Lorenzato (1995) can be written in this form as well, + but assumes approximate lognormal distribution of errors for high values of y. } \usage{ -sigma_rl(y, sigma_low, rsd_high) +sigma_twocomp(y, sigma_low, rsd_high) } \arguments{ \item{y}{ The magnitude of the observed value } @@ -21,6 +25,10 @@ sigma_rl(y, sigma_low, rsd_high) The standard deviation of the response variable. } \references{ - Rocke, David M. und Lorenzato, Stefan (1995) A two-component model for + Werner, Mario, Brooks, Samuel H., and Knott, Lancaster B. (1978) + Additive, Multiplicative, and Mixed Analytical Errors. Clinical Chemistry + 24(11), 1895-1898. + + Rocke, David M. and Lorenzato, Stefan (1995) A two-component model for measurement error in analytical chemistry. Technometrics 37(2), 176-184. } diff --git a/man/synthetic_data_for_UBA.Rd b/man/synthetic_data_for_UBA.Rd index a9df9767..f9d3c77b 100644 --- a/man/synthetic_data_for_UBA.Rd +++ b/man/synthetic_data_for_UBA.Rd @@ -13,7 +13,10 @@ Variance component 'b' is also based on a normal distribution, but with a standard deviation of 7.
Variance component 'c' is based on the error model from Rocke and Lorenzato (1995), with the
minimum standard deviation (for small y values) of 0.5, and a proportionality constant of 0.07
- for the increase of the standard deviation with y.
+ for the increase of the standard deviation with y. Note that this is a simplified version
+ of the error model proposed by Rocke and Lorenzato (1995), as in their model the error of the
+ measured values approximates lognormal distribution for high values, whereas we are using
+ normally distributed error components all along.
Initial concentrations for metabolites and all values where adding the variance component resulted
in a value below the assumed limit of detection of 0.1 were set to \code{NA}.
|