NAME
dbproject_covariance- compute confidence-ellipse from covariance matrix
SYNOPSIS
dbproject_covariance [-v] [-c conf] dbname
SUPPORT
Contributed code: NO BRTT support.
THIS PIECE OF SOFTWARE WAS CONTRIBUTED BY THE ANTELOPE USER COMMUNITY. BRTT DISCLAIMS ALL OWNERSHIP, LIABILITY, AND SUPPORT FOR THIS PIECE OF SOFTWARE.
FOR HELP WITH THIS PIECE OF SOFTWARE, PLEASE CONTACT THE CONTRIBUTING AUTHOR.
DESCRIPTION
dbproject_covariance reads the covariance matrix out of the origerr
table of a CSS3.0 or related database. A confidence ellipse is then
calculated (smajax, sminax, and strike, as well as sdepth and stime)
based on a Chi-squared model and the specified confidence level. If
no confidence level is specified,
dbproject_covariance uses 0.683,
i.e. 1 sigma. If any of smajax, sminax, strike, sdepth, and stime are
non-null, the origerr row is skipped. If any of the covariance-matrix
elements are null, the origerr row is skipped.
dbproject_covariance
also uses ndef and dtype out of the origin table to determine the number
of degrees of freedom of the solution.
When the projected confidence region has been computed, the origerr
table is updated with the new smajax, sminax, sdepth, stime, strike, and
conf values.
OPTIONS
-
-c conf
Confidence level, from 0 to 1
-
-v
Verbose
EXAMPLE
% dbproject_covariance -v testdb
orid 1 ndef 41 dtype f conf 0.683 chi2 40.54
smajax 0.95 sminax 0.55 strike 149.5
sdepth 3.61 stime 0.30
orid 2 ndef 56 dtype f conf 0.683 chi2 56.31
smajax 3.53 sminax 2.24 strike 18.9
sdepth 6.30 stime 0.53
orid 3 ndef 21 dtype f conf 0.683 chi2 19.20
smajax 9.10 sminax 4.01 strike 114.0
sdepth 9.38 stime 0.93
orid 4 ndef 60 dtype f conf 0.683 chi2 60.49
smajax 0.64 sminax 0.58 strike 98.6
sdepth 5.33 stime 0.33
orid 5 ndef 41 dtype f conf 0.683 chi2 40.54
smajax 1.59 sminax 0.60 strike 121.6
sdepth 2.87 stime 0.20
%
SEE ALSO
dbgenloc(1)
BUGS AND CAVEATS
The number of degrees of freedom is based on ndef and dtype. If array
data are included (azimuth and slowness information in addition to
arrival-time), more than one degree of freedom is added per incoming
phase, so using the number of defining phases is not exact. To
do this properly, the timedef, azdef, and slodef fields would have to be
examined in the assoc table. This has not been done in the current program.
There is another way to calculate the confidence level, namely with
an F-distribution statistical model rather than the chi-squared. In
principle this would be possible however it would probably require
a recomputation of residuals from the assoc table.
AUTHOR
Kent Lindquist
Antelope User Group Contributed Software