[8e1f3a]: mle_estimate.m  Maximize  Restore  History

Download this file

75 lines (66 with data), 3.0 kB

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
## Copyright (C) 2003,2004,2005 Michael Creel <michael.creel@uab.es>
##
## This program is free software; you can redistribute it and/or modify
## it under the terms of the GNU General Public License as published by
## the Free Software Foundation; either version 2 of the License, or
## (at your option) any later version.
##
## This program is distributed in the hope that it will be useful,
## but WITHOUT ANY WARRANTY; without even the implied warranty of
## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
## GNU General Public License for more details.
##
## You should have received a copy of the GNU General Public License
## along with this program; if not, write to the Free Software
## Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
# usage:
# [theta, obj_value, conv, iters] = mle_estimate(theta, data, model, modelargs, control, nslaves)
#
# inputs:
# theta: column vector of model parameters
# data: data matrix
# model: name of function that computes log-likelihood
# modelargs: (cell) additional inputs needed by model. May be empty ("")
# control: (optional) BFGS or SA controls (see bfgsmin and samin). May be empty ("").
# nslaves: (optional) number of slaves if executed in parallel (requires MPITB)
#
# outputs:
# theta: ML estimated value of parameters
# obj_value: the value of the log likelihood function at ML estimate
# conv: return code from bfgsmin (1 means success, see bfgsmin for details)
# iters: number of BFGS iteration used
#
# please see mle_example.m for examples of how to use this
function [theta, obj_value, convergence, iters] = mle_estimate(theta, data, model, modelargs, control, nslaves)
if nargin < 3
error("mle_estimate: 3 arguments required");
endif
if nargin < 4 modelargs = {}; endif # create placeholder if not used
if !iscell(modelargs) modelargs = {}; endif # default controls if receive placeholder
if nargin < 5 control = {Inf,0,1,1}; endif # default controls and method
if !iscell(control) control = {Inf,0,1,1}; endif # default controls if receive placeholder
if nargin < 6 nslaves = 0; endif
if nslaves > 0
global NSLAVES PARALLEL NEWORLD TAG;
LAM_Init(nslaves);
# Send the data to all nodes
NumCmds_Send({"data", "model", "modelargs"}, {data, model, modelargs});
endif
# bfgs or sa?
if (size(control,1)*size(control,2) == 0) # use default bfgs if no control
control = {Inf,0,1,1};
method = "bfgs";
elseif (size(control,1)*size(control,2) < 11)
method = "bfgs";
else method = "sa";
endif
if strcmp(method, "bfgs")
[theta, obj_value, convergence, iters] = bfgsmin("mle_obj", {theta, data, model, modelargs, nslaves}, control);
elseif strcmp(method, "sa")
[theta, obj_value, convergence] = samin("mle_obj", {theta, data, model, modelargs, nslaves}, control);
endif
if nslaves > 0
LAM_Finalize;
endif # cleanup
obj_value = - obj_value; # recover from minimization rather than maximization
endfunction

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:





No, thanks