loc.est-class {metRology} | R Documentation |
The location estimate class
Description
The location estimate class contains output from a variety of estimators used in the metRology package.
A print method is provided.
Usage
## S3 method for class 'loc.est'
print(x, ...)
Arguments
x |
An object of class ‘loc.est’ |
... |
Parameters passed to other functions. Currently unused. |
Details
An object of class ‘loc.est’ is a list containing
- x
Scalar estimate of location
- u
Standard uncertainty (usually equivalent to standard error) of the location estimate.
- df
Degrees of freedom associated with the location estimate (may be NA)
- xi
Numeric vector of individual values contributing to the estimate
- ui
Numeric vector of uncertainties initially associated with
xi
.- dfi
Numeric vector of degrees of freedom associated with
ui
.- u.eff
Numeric vector of ‘effective uncertainties’ in
xi
after any additional terms or adjustments are added (see below).- w
Numeric vector of weights associated with
xi
(see below).- method
Character string describing the method used to obtain the estimate.
- method.details
An optional list of additional details provided by the particular method used.
The ‘effective uncertainties’ u.eff
arise from some estimation methods (for example,
Mandel-Paule). These typically involve either the estimation of an additional variance term,
a scale adjustment to the output value uncertainty or (for example in the case of the
arithmetic mean) replacement of the initial individual uncertainties with some single
estimate based on the dispersion of values. These adjustments are usually equivalent to
replacing the estimator used with a weighted mean using weights 1/u_{eff}^2
.
The weight vector w
is not equivalent to 1/u_{eff}^2
. Rather, it
gives the ratio of prior weights 1/u_{eff}^2
to posterior weights, which combine
prior weights with some additional weighting. Posterior weights arise in particular when using
robust estimators, and are generally 1 otherwise. The returned location estimate in such cases
can be calculated as sum(w*x/(u^2))/sum(w/(u^2)))
.
method.details
is an optional list that may contain anything from a short
summary of a scale factor or additional variance to a complete object (e.g. an rlm
object) returned by the function used to calculate the estimate.
Value
The print method is called for its side effect; no value is returned.
Author(s)
S. L. R. Ellison s.ellison@lgc.co.uk
References
None, yet
See Also
Examples
## Cd heat of vapourisation example (see ?mpaule)
x2<-c(27.044, 26.022, 26.340, 26.787, 26.796)
v<-c(3, 76, 464, 3, 14)*1e-3
mp<-mpaule(x2, sqrt(v))
print(mp)