Module functions to display contours of statistics as a function of parameter values, and to retrieve arrays of values and statistics
Struct_Type run_regunc(Array_Type)
Struct_Type run_regproj(Array_Type)
Argument:
(1) An array of two strings representing two thawed model parameters.
These functions initiate the
REGION-UNCERTAINTY and
REGION-PROJECTION
parameter estimation methods respectively.
The chosen method is run using
the most recently fit datasets, which are automatically determined and
hence are not function arguments.
When done, each returns a structure,
which are the same as those returned by
get_regunc and
get_regproj.
Fit a dataset; get information about chi-square
as a function of power-law amplitude p.ampl and
slope p.gamma
sherpa> () = load_ascii(1,"spec.dat")
sherpa> () = create_model("POLY","p")
sherpa> set_thawed(["p.c1","p.c2","p.c3"])
sherpa> () = set_source_expr(1,"p")
sherpa> () = run_fit
LVMQT: V2.0
LVMQT: initial statistic value = 82.2297
LVMQT: final statistic value = 62.2247 at iteration 3
p.c0 61.4774
p.c1 -0.380228
p.c2 0.00993229
p.c3 -7.01741e-05
sherpa> list_regproj()
Parameter Current Default Description
----------------------------------------------------------------------
fast 1 1 Switch to LM/simplex: 0(n)/1(y)
expfac 3 3 Expansion factor for grid
arange 1 1 Auto-range: 0(n)/1(y)
min [0,0] [0,0] Minimum values, each axis
max [0,0] [0,0] Maximum values, each axis
log [0,0] [0,0] Log-spacing: 0(n)/1(y), each axis
nloop [10,10] [10,10] Number of grid points, each axis
sigma [1,2,3] [1,2,3] Number of sigma, each contour
sherpa> sherpa.regproj.sigma = [1.6,2.6]
sherpa> regproj = run_regproj(["p.c0","p.c1"])
[...plot displayed...]
sherpa> print(regproj)
x0 = Float_Type[100]
x1 = Float_Type[100]
y = Float_Type[100]
levels = Float_Type[2]
name = String_Type[2]
bfit = Double_Type[2]
config = sherpa_VisParEst_State
sherpa> printarr(regproj.x0,3)
61.3661
61.3661
61.3661
sherpa> printarr(regproj.x1,3)
-1.79518
-1.48075
-1.16631
sherpa> printarr(regproj.y,3)
154.651
118.31
90.9931
The third-to-last and second-to-last function calls
cause the first three values of
the p.c0 and p.c1 grid axes to
be displayed (with p.c0 repeated because a 2-D
grid is being computed).
The last function call displays
the best-fit statistic given the p.c0 and p.c1
values.
The run_regunc() and run_regproj() commands no longer place two arrays on the stack. This means that you can use
retval = run_regproj(parameter_names); rather than having to say something like
(,,retval) = run_regproj(parameter_names);
The structures returned by these functions contain additional fields: levels, name, bfit, and config. These fields contain information on the statistic value for each contour, the name of the parameters, their best-fit values, and the values used by the "region" command to calculate the x0, x1, and y values.
These functions will only run after the dataset has been fitted; i.e run_fit() called in the same session. This is unlike the Sherpa versions of these commands, which have been updated in CIAO 3.2 to not require the inital fit.
See the Sherpa bug pages online for an up-to-date listing of known bugs.
- sherpa
-
berrors,
bsyserrors,
compute_errors,
compute_statistic,
covariance,
errors,
ftest,
get_paramest,
get_paramestint,
get_paramestlim,
get_paramestreg,
goodness,
interval-projection,
interval-uncertainty,
list_paramest,
mlr,
projection,
region-projection,
region-uncertainty,
restore_paramest,
run_paramest,
run_paramestint,
run_paramestlim,
set_errors,
set_syserrors,
staterrors,
syserrors,
uncertainty
|
![[CIAO Logo]](../imgs/ciao_logo_navbar.gif){kind=logo}