get_lrt_results

Context: plotting

Synopsis

Access the simulation results of the likelihood ratio test.

Syntax

get_lrt_results()

The get_lrt_results() command returns the likelihood ratio test results computed by the plot_lrt() command, comparing fits of the null model to fits of the alternative model using faked data with Poisson noise. The likelihood ratio based on the observed data is returned, along with the p-value, used to reject or accept the null model.

The following attributes are returned by get_lrt_results():

samples - parameter samples from all iterations, in the order of the parameter list (as many as samples as specified in the 'num' parameter of plot_lrt() are returned; default is 500)
stats - calculated fit statistics in all iterations (first column null, alt second)
ratios - calculated likelihood ratios from all iterations
null - fit statistic of null on real data
alt - fit statistic of alt on real data
lr - likelihood ratio on real data
ppp - p-value

Example 1

sherpa> plot_lrt(powlaw1d.p1, p1+gauss1d.g1)
sherpa> print get_lrt_results()

samples = [[  8.5623e-01   5.6209e+03]
 [  8.5391e-01   5.5819e+03]
 [  8.5695e-01   5.7627e+03]
 ..., 
 [  8.6336e-01   5.9058e+03]
 [  8.5459e-01   5.6061e+03]
 [  8.3418e-01   4.9820e+03]]
stats   = [[ 466.1517  466.1517]
 [ 500.005   500.005 ]
 [ 474.0595  474.0595]
 ..., 
 [ 424.5051  424.5051]
 [ 453.3375  453.3375]
 [ 463.3019  463.3019]]
ratios  = [-0. -0. -0. ..., -0. -0. -0.]
null    = 3958.4582610638136
alt     = 3958.4582610638136
lr      = -0.0
ppp     = 0.054

Call the get_lrt_results() command with no arguments to print to the screen the results of the likelihood ratio test performed by the plot_lrt() command. View the array of ratios comparing fits to simulated data done with a simple (null) power-law model versus the more-complex (alternative) power-law plus Gaussian model, along with the likelihood ratio and p-value based on the observed data.

Example 2

lrt = get_lrt_results()

Store the data returned by get_lrt_results() to variable "lrt", in order to access and manipulate individual attributes within a script.

sherpa> res.ratios
            array([-0., -0., -0., ..., -0., -0., -0.])

sherpa> res.ppp
            0.054

sherpa> res.stats
            
array([[ 466.1517,  466.1517],
       [ 500.005 ,  500.005 ],
       [ 474.0595,  474.0595],
       ..., 
       [ 424.5051,  424.5051],
       [ 453.3375,  453.3375],
       [ 463.3019,  463.3019]])


sherpa> len(res.stats)
            500

sherpa> res.stats[46]
            array([ 532.5428,  532.5428])

Bugs

See the bugs pages on the Sherpa website for an up-to-date listing of known bugs.