this is for holding javascript data
Antonino Ingargiola edited background-estimation.tex
about 8 years ago
Commit id: e149a6539594a91914f6476625496833e7b01e73
deletions | additions
diff --git a/background-estimation.tex b/background-estimation.tex
index 9483d69..d857b2d 100644
--- a/background-estimation.tex
+++ b/background-estimation.tex
...
\subsubsection{Error metric and optimal threshold}
The functions
that used to fit the background
provide also
return an estimation of a goodness-of-fit estimator
computed on the
quality basis of the
fit, computed as the distance empirical distribution function (EDF)~\cite{Stephens1974,Parr1980}.
The ``distance'' between the
empirical
\href{http://en.wikipedia.org/wiki/Cumulative\_distribution\_function}{cumulative EDF and the theoretical (i.e. exponential) cumulative distribution
function}
(CDF) represent and
fitted CDF. indicator of the quality of fit.
Two different distance metrics can be
returned. returned by the background fitting functions.
The first is the
\href{http://en.wikipedia.org/wiki/Kolmogorov\%E2\%80\%93Smirnov\_test}{Kolgomorov-Smirnov}
statistics (the Kolgomorov-Smirnov statistics, which uses the maximum of the difference
between the
empirical and the
fitted CDF) EDF and the
theoretical distribution. The second is the
\href{http://en.wikipedia.org/wiki/Cram\%C3\%A9r\%E2\%80\%93von\_Mises\_criterion}{Cramér Cramér von
Mises} Mises
statistics corresponding to the integral of the squared residuals
(see the code
\href{https://github.com/tritemio/FRETBursts/blob/master/fretbursts/background.py#L41}{here}). for more details,
\href{https://github.com/tritemio/FRETBursts/blob/master/fretbursts/background.py#L41}{link}).
In principle, the optimal
interphoton-delay threshold
minimizes will minimize
the error metric. This approach is implemented by the function
\href{http://fretbursts.readthedocs.org/en/latest/plugins.html#fretbursts.burstlib\_ext.calc\_bg\_brute}{calc\_bg\_brute}
in the
\href{http://fretbursts.readthedocs.org/en/latest/plugins.html}{burstlib\_ext module} and can be used
in cases when \verb|calc_bg_brute|
(\href{http://fretbursts.readthedocs.org/en/latest/plugins.html#fretbursts.burstlib\_ext.calc\_bg\_brute}{link}) which performs a
user wants brute-force search in order to
extract find the
most accurate estimation optimal threshold.
This level of
sofistication in estimation the background
rates. rates is not
necessary under typical circumstances, as the difference between an optimal threshold
and a manually (or euristically) choosen one is negligible in most practical cases.