IPP Software Navigation Tools IPP Links Communication Pan-STARRS Links

Ignore:
Timestamp:
May 24, 2004, 10:14:06 AM (22 years ago)
Author:
eugene
Message:

added references to various external documents
minor changes to robust stats (bin size from clipped sigma)

File:
1 edited

Legend:

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

    r756 r758  
    1 %%% $Id: psLibADD.tex,v 1.16 2004-05-24 19:27:53 eugene Exp $
     1%%% $Id: psLibADD.tex,v 1.17 2004-05-24 20:14:06 eugene Exp $
    22\documentclass[panstarrs]{panstarrs}
    33
     
    4242SLALIB Positional Astronomy Library & http://star-www.rl.ac.uk/star/docs/sun67.htx/sun67.html \\ \hline
    4343Numerical Recipes (NR)              & \\ \hline
    44 Knuth                               & \\ \hline
    45 Sedgewick                           & \\ \hline
     44Knuth, D.E.                         & Sorting and Searching; The Art of Computer Programming \\ \hline
     45Sedgewick, R.                       & Algorithms, Ch. 8 \\ \hline
    4646Sorting Summary                     & {\tt http://www.iti.fh-flensburg.de/lang/algorithmen/sortieren/algoen.htm } \\ \hline
    4747GSL                                 & \\ \hline
     
    231231valued, the natural bin size is an integer.  Otherwise, the bin should
    232232be a fraction of an estimate of the standard deviation.  Use the
    233 sample upper and lower quartiles to determine an estimate of the
    234 standard deviation: $\sigma_e = (U_{\frac{1}{4}} - L_{\frac{1}{4}}) /
    235 1.34$.  The bin size shall be set at $\sigma_e / 10$.  The remaining
    236 steps of the algorithm are as follows:
     233$3\sigma$ clipped standard deviation as an estimator of the standard
     234deviation.  The bin size shall be set at $\sigma_e / 10$.  The
     235remaining steps of the algorithm are as follows:
    237236
    238237\begin{itemize}
     
    242241\item Find the bin with the peak value in the range $L_{\frac{1}{4}}$
    243242  to $U_{\frac{1}{4}}$; this is the robust mode, $\mbox{mode}_r$. 
    244 \item Determine $dL = (U_{\frac{1}{4}} - L_{\frac{1}{4}}) / 8$.
     243\item Determine $dL = (U_{\frac{1}{4}} - L_{\frac{1}{4}}) / 4$.
    245244\item Fit a Gaussian to the bins in the range $\mbox{mode}_r - dL$ to
    246245  $\mbox{mode}_r + dL$.
     
    253252percentile value and its two neighbors.  Fit a quadratic to these
    254253three points.  The robust median value is the coordinate of the
    255 quadratic which returns the 50\% value.
     254quadratic which returns the 50\% value.  For the upper and lower
     255quartile points, the same process should be used, choosing the three
     256bins in the vicinity of the upper and lower quartile points.
    256257
    257258
Note: See TracChangeset for help on using the changeset viewer.