IPP Software Navigation Tools IPP Links Communication Pan-STARRS Links

Ignore:
Timestamp:
Dec 21, 2004, 11:37:08 AM (22 years ago)
Author:
Paul Price
Message:

Adding extra information on histograms, including an example of
histograms in the presence of errors.

File:
1 edited

Legend:

Unmodified
Added
Removed
  • trunk/doc/pslib/psLibADD.tex

    r2641 r2779  
    1 %%% $Id: psLibADD.tex,v 1.55 2004-12-06 21:35:43 price Exp $
     1%%% $Id: psLibADD.tex,v 1.56 2004-12-21 21:37:08 price Exp $
    22\documentclass[panstarrs]{panstarrs}
    33
     
    304304contribute a fraction of a value, equivalent to the weight.  In the
    305305interests of speed, a boxcar PDF may be used to represent each input
    306 value (as opposed to a Gaussian), where the width is equal to 2.35
    307 times the error (i.e., the full width at half maximum) and each input
    308 value contributes constant area.  Then the mean, median, mode,
    309 standard deviation and quartiles are estimated in the same manner as
    310 above.
     306value (as opposed to a Gaussian), where the boxcar width is equal to
     307$2 \sqrt{2 \ln 2}$ times the error and each input value contributes
     308constant area.  Then the mean, median, mode, standard deviation and
     309quartiles are estimated in the same manner as above.
    311310
    312311\paragraph{Histograms}
     
    315314input values, the approach described above for the robust statistics
    316315is used (i.e., the histograms become probability density functions).
     316
     317An example may help here.  Say we have our histogram bounds being 0,
     3181, 2, 3, 4, 5; and our value is $2.5 \pm 0.5$.  Then, the width of the
     319contribution is $0.5 \times 2.35... \approx 1.175$.  Half the width is
     3200.5875, so we will treat this value as a boxcar from $2.5 - 0.5875$ to
     321$2.5 + 0.5875$.
     322
     323Consequently, the bins 0 to 1 and 4 to 5 get no value, because none of
     324the boxcar overlaps.  The bin 1 to 2 gets 0.0875, because that's the
     325fraction of the boxcar that overlaps with it; same thing with the bin
     3263 to 4.  The bin 2 to 3 gets 0.825 because that's the fraction of the
     327boxcar that overlaps with it.  So the single value $2.5 \pm 0.5$ makes
     328the following histogram:
     329
     330\begin{tabular}{lr}
     331Bin & Value \\ \hline
     3320--1 & 0 \\
     3331--2 & 0.0875 \\
     3342--3 & 0.8250 \\
     3353--4 & 0.0875 \\
     3364--5 & 0 \\
     337\end{tabular}
     338
     339Note that the total adds to one --- the number of values added.
    317340
    318341\subsubsection{Matrix Operations}
Note: See TracChangeset for help on using the changeset viewer.