[Met_help] [rt.rap.ucar.edu #40817] History for Stat-Analysis Tool Question

RAL HelpDesk {for Paul Oldenburg} met_help at ucar.edu
Thu Sep 16 10:06:48 MDT 2010


----------------------------------------------------------------
  Initial Request
----------------------------------------------------------------

MET Help,

Can Stat-Analysis add together (not average!) multiple contingency tables and compute stats 
on those totals?  For example, if I have 10 forecasts valid at the same forecast hour and n = 
65536 (256x256) for each one, how can I add them all together to get n = 655360 (with 
associated a, b, c, and d values from the contingency table) and then compute the statistics 
on those values without averaging first?  Hopefully I explained my question well enough, but 
if not, let me know.

Thanks,
Derek Stratman


----------------------------------------------------------------
  Complete Ticket History
----------------------------------------------------------------

Subject: Re: [rt.rap.ucar.edu #40817] Stat-Analysis Tool Question
From: Paul Oldenburg
Time: Thu Sep 16 09:41:14 2010

Derek,

One of the fundamental uses of stat_analysis is to add contingency
table counts from different cases and recompute
counts and statistics.  There are a few examples of how to do this
below.  In order to filter the stat_analysis input so
that it only grabs the desired CTC data, use the filtering options.
In your case, stat_analysis unfortunately does not
support the fcst_valid_hour filtering option, so you might have to get
creative using a combination of the
fcst_init_hour and the fcst_lead filtering options.  You may consider
using stat_analysis to create intermediate .stat
files using the filter job.  The good news is that we just decided to
add the fcst_valid_hour and obs_valid_hour filter
criteria to stat_analysis, so those will be available in METv3.0.  If
you have any questions, please let me know.

Thanks,

Paul


$ stat_analysis -lookin data/ -job aggregate -line_type CTC
JOB_LIST:       -job aggregate -line_type CTC
COL_NAME: TOTAL  FY_OY FY_ON FN_OY FN_ON
     CTC: 178442 39939 12022 11892 114589

$ stat_analysis -lookin data/ -job aggregate_stat -line_type CTC
-out_line_type CTS
JOB_LIST:       -job aggregate_stat -line_type CTC -out_line_type CTS
-out_alpha 0.050000
COL_NAME: TOTAL  BASER   BASER_NCL BASER_NCU BASER_BCL BASER_BCU FMEAN
FMEAN_NCL FMEAN_NCU FMEAN_BCL FMEAN_BCU ...
     CTS: 178442 0.29046 0.28836   0.29257   NA        NA
0.29119 0.28909   0.29330   NA        NA        ...

$ stat_analysis -lookin data/ -job summary -line_type CTS -column GSS
GSL_RNG_TYPE=mt19937
GSL_RNG_SEED=1194693196
JOB_LIST:       -job summary -line_type CTS -column GSS -out_alpha
0.050000
COL_NAME: TOTAL MEAN    MEAN_NCL MEAN_NCU MEAN_BCL MEAN_BCU STDEV
STDEV_BCL STDEV_BCU MIN      P10     P25     ...
 SUMMARY: 80    0.23721 0.20049  0.27394  0.20280  0.27243  0.16759
0.15109   0.18092   -0.01212 0.01925 0.08347 ...




RAL HelpDesk {for Derek.Stratman} wrote:
> Thu Sep 16 08:56:46 2010: Request 40817 was acted upon.
> Transaction: Ticket created by Derek.Stratman at noaa.gov
>        Queue: met_help
>      Subject: Stat-Analysis Tool Question
>        Owner: Nobody
>   Requestors: Derek.Stratman at noaa.gov
>       Status: new
>  Ticket <URL:
https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=40817 >
>
>
> MET Help,
>
> Can Stat-Analysis add together (not average!) multiple contingency
tables and compute stats
> on those totals?  For example, if I have 10 forecasts valid at the
same forecast hour and n =
> 65536 (256x256) for each one, how can I add them all together to get
n = 655360 (with
> associated a, b, c, and d values from the contingency table) and
then compute the statistics
> on those values without averaging first?  Hopefully I explained my
question well enough, but
> if not, let me know.
>
> Thanks,
> Derek Stratman


------------------------------------------------
Subject: Re: [rt.rap.ucar.edu #40817] Stat-Analysis Tool Question
From: Derek.Stratman
Time: Thu Sep 16 10:05:39 2010

Paul,

Thanks for your help.  For some reason I thought Stat-Analysis took
the stats and simply
averaged them, but that's clearly not the case.  My situation will be
rather easy then.

Thanks again for your clarification,

-Derek

----- Original Message -----
From: RAL HelpDesk {for Paul Oldenburg} <met_help at ucar.edu>
Date: Thursday, September 16, 2010 10:41 am
Subject: Re: [rt.rap.ucar.edu #40817] Stat-Analysis Tool Question
> Derek,
>
> One of the fundamental uses of stat_analysis is to add contingency
> table counts from different cases and recompute
> counts and statistics.  There are a few examples of how to do this
> below.  In order to filter the stat_analysis input so
> that it only grabs the desired CTC data, use the filtering options.
> In your case, stat_analysis unfortunately does not
> support the fcst_valid_hour filtering option, so you might have to
> get creative using a combination of the
> fcst_init_hour and the fcst_lead filtering options.  You may
> consider using stat_analysis to create intermediate .stat
> files using the filter job.  The good news is that we just decided
> to add the fcst_valid_hour and obs_valid_hour filter
> criteria to stat_analysis, so those will be available in METv3.0.
> If you have any questions, please let me know.
>
> Thanks,
>
> Paul
>
>
> $ stat_analysis -lookin data/ -job aggregate -line_type CTC
> JOB_LIST:       -job aggregate -line_type CTC
> COL_NAME: TOTAL  FY_OY FY_ON FN_OY FN_ON
>     CTC: 178442 39939 12022 11892 114589
>
> $ stat_analysis -lookin data/ -job aggregate_stat -line_type CTC -
> out_line_type CTS
> JOB_LIST:       -job aggregate_stat -line_type CTC -out_line_type
> CTS -out_alpha 0.050000
> COL_NAME: TOTAL  BASER   BASER_NCL BASER_NCU BASER_BCL BASER_BCU
> FMEAN   FMEAN_NCL FMEAN_NCU FMEAN_BCL FMEAN_BCU ...
>     CTS: 178442 0.29046 0.28836   0.29257   NA        NA
> 0.29119 0.28909   0.29330   NA        NA        ...
>
> $ stat_analysis -lookin data/ -job summary -line_type CTS -column
GSS
> GSL_RNG_TYPE=mt19937
> GSL_RNG_SEED=1194693196
> JOB_LIST:       -job summary -line_type CTS -column GSS -out_alpha
> 0.050000COL_NAME: TOTAL MEAN    MEAN_NCL MEAN_NCU MEAN_BCL MEAN_BCU
> STDEV   STDEV_BCL STDEV_BCU MIN      P10     P25     ...
> SUMMARY: 80    0.23721 0.20049  0.27394  0.20280  0.27243  0.16759
> 0.15109   0.18092   -0.01212 0.01925 0.08347 ...
>
>
>
>
> RAL HelpDesk {for Derek.Stratman} wrote:
> > Thu Sep 16 08:56:46 2010: Request 40817 was acted upon.
> > Transaction: Ticket created by Derek.Stratman at noaa.gov
> >        Queue: met_help
> >      Subject: Stat-Analysis Tool Question
> >        Owner: Nobody
> >   Requestors: Derek.Stratman at noaa.gov
> >       Status: new
> >  Ticket <URL:
> https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=40817 >
> >
> >
> > MET Help,
> >
> > Can Stat-Analysis add together (not average!) multiple
> contingency tables and compute stats
> > on those totals?  For example, if I have 10 forecasts valid at
> the same forecast hour and n =
> > 65536 (256x256) for each one, how can I add them all together to
> get n = 655360 (with
> > associated a, b, c, and d values from the contingency table) and
> then compute the statistics
> > on those values without averaging first?  Hopefully I explained
> my question well enough, but
> > if not, let me know.
> >
> > Thanks,
> > Derek Stratman
>
>
>

------------------------------------------------


More information about the Met_help mailing list