r16161 - /branches/uf_redesign/user_functions/monte_carlo.py -- May 09, 2012 - 17:03

Author: bugman
Date: Wed May  9 17:03:26 2012
New Revision: 16161

URL: http://svn.gna.org/viewcvs/relax?rev=16161&view=rev
Log:
Eliminated the 'prune' argument to the monte_carlo.error_analysis user 
function.

This will never be used again, and might tempt some users into changing the 
value and damaging their
analysis!


Modified:
    branches/uf_redesign/user_functions/monte_carlo.py

Modified: branches/uf_redesign/user_functions/monte_carlo.py
URL: 
http://svn.gna.org/viewcvs/relax/branches/uf_redesign/user_functions/monte_carlo.py?rev=16161&r1=16160&r2=16161&view=diff
==============================================================================
--- branches/uf_redesign/user_functions/monte_carlo.py (original)
+++ branches/uf_redesign/user_functions/monte_carlo.py Wed May  9 17:03:26 
2012
@@ -116,17 +116,8 @@
 uf = uf_info.add_uf('monte_carlo.error_analysis')
 uf.title = "Calculate parameter errors from the Monte Carlo simulations."
 uf.title_short = "Error calculation."
-uf.add_keyarg(
-    name = "prune",
-    default = 0.0,
-    py_type = "num",
-    desc_short = "pruning quantity",
-    desc = "Legacy argument corresponding to 'trim' in Art Palmer's 
Modelfree program.  The default value of 0.0 corresponds to no pruning of the 
data."
-)
 uf.desc = """
 Parameter errors are calculated as the standard deviation of the 
distribution of parameter values.  This function should never be used if 
parameter values are obtained by minimisation and the simulation data are 
generated using the method 'direct'.  The reason is because only true Monte 
Carlo simulations can give the true parameter errors.
-
-The prune argument is legacy code which corresponds to the 'trim' option in 
Art Palmer's Modelfree program.  To remove failed simulations, the eliminate 
function should be used prior to this function.  Eliminating the simulations 
specifically identifies and removes the failed simulations whereas the prune 
argument will only, in a few cases, positively identify failed simulations 
but only if severe parameter limits have been imposed.  Most failed models 
will pass through the pruning process and hence cause a catastrophic increase 
in the parameter errors.  If the argument must be used, the following must be 
taken into account. If the values or parameters are calculated rather than 
minimised, the prune argument must be set to zero.  The value of this 
argument is proportional to the number of simulations removed prior to error 
calculation.  If prune is set to 0.0, all simulations are used for 
calculating errors, whereas a value of 1.0 excludes all data.  In almost all 
cases prune must be set to zero, any value greater than zero will result in 
an underestimation of the error values.  If a value is supplied, the lower 
and upper tails of the distribution of chi-squared values will be excluded 
from the error calculation.
 """
 uf.additional = [monte_carlo_desc]
 uf.backend = monte_carlo.error_analysis