Next: Normalization uncertainty
Up: Building the covariance matrix
Previous: Building the covariance matrix
Contents
Let
be the
results
of independent measurements
and
the (diagonal) covariance matrix.
Let us assume that they are all affected by the same calibration
constant , having a standard uncertainty .
The corrected results are then
.
We can assume, for
simplicity, that the most probable value of is 0, i.e.
the detector is well calibrated.
One has to
consider the calibration constant as
the physical quantity , whose best estimate is
.
A term
must be added to the
covariance matrix.
The covariance matrix of the corrected results is given by the
transformation
|
(6.25) |
where
.
The elements of are given by
|
(6.26) |
In this case we get
|
|
|
(6.27) |
Cov |
|
|
(6.28) |
|
|
|
(6.29) |
|
|
|
(6.30) |
reobtaining the results of Section .
The total uncertainty on the single measurement is given by the
combination in quadrature of the individual and the common
standard uncertainties, and all the covariances are equal to
.
To verify, in a simple case, that the result is reasonable,
let us consider only two independent quantities and ,
and a calibration constant , having
an expected value equal to zero. From these we can calculate
the correlated quantities and and finally their
sum (
) and difference (
). The results are
It follows that
as intuitively expected.
Next: Normalization uncertainty
Up: Building the covariance matrix
Previous: Building the covariance matrix
Contents
Giulio D'Agostini
2003-05-15