Mercurial > hg > ltpda
view m-toolbox/test/test_matrix_linlsqsvd.m @ 30:317b5f447f3e database-connection-manager
Update workspaceBrowser
author | Daniele Nicolodi <nicolodi@science.unitn.it> |
---|---|
date | Mon, 05 Dec 2011 16:20:06 +0100 |
parents | f0afece42f48 |
children |
line wrap: on
line source
% TEST_MATRIX_LINLSQSVD tests the linlsqsvd method of the AO class. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % L Ferraioli 10-11-2010 % % $Id: test_matrix_linlsqsvd.m,v 1.1 2011/02/18 17:07:35 luigi Exp $ % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %% 1) Determine the coefficients of a linear combination of noises and %% comapre with lscov: % % Make some data fs = 10; nsecs = 10; B1 = ao(plist('tsfcn', 'randn(size(t))', 'fs', fs, 'nsecs', nsecs, 'yunits', 'T')); B1.setName; B2 = ao(plist('tsfcn', 'randn(size(t))', 'fs', fs, 'nsecs', nsecs, 'yunits', 'T')); B2.setName; B3 = ao(plist('tsfcn', 'randn(size(t))', 'fs', fs, 'nsecs', nsecs, 'yunits', 'T')); B3.setName; B4 = ao(plist('tsfcn', 'randn(size(t))', 'fs', fs, 'nsecs', nsecs, 'yunits', 'T')); B4.setName; C1 = matrix(B1,B2,plist('shape',[2,1])); C1.setName; C2 = matrix(B3,B4,plist('shape',[2,1])); C2.setName; C = matrix([B1 B3;B2 B4]); C.setName; n1 = ao(plist('tsfcn', 'randn(size(t))', 'fs', fs, 'nsecs', nsecs, 'yunits', 'm')); n2 = ao(plist('tsfcn', 'randn(size(t))', 'fs', fs, 'nsecs', nsecs, 'yunits', 'm')); n = matrix(n1,n2,plist('shape',[2,1])); n.setName; a = [ao(1,plist('yunits','m/T')) ao(2,plist('yunits','m/T'))]; A = matrix(a,plist('shape',[2,1])); % assign output values y = C*A + n; %%% Get a fit with linlsqsvd pobj1 = linlsqsvd(C1, C2, y); % combine results for ii=1:numel(pobj1.y) prs(ii) = ao(cdata(pobj1.y(ii))); prs(ii).setYunits(pobj1.yunits(ii)); end Pars = matrix(prs,plist('shape',[numel(prs),1])); yfit1 = C*Pars; %%% do linear combination: using eval yfit2 = pobj1.eval; % Plot (compare data with fit) iplot(y.objs(1), yfit1.objs(1), yfit2.objs(1)) iplot(y.objs(2), yfit1.objs(2), yfit2.objs(2)) %% 2) Determine the coefficients of a linear combination of noises: % % Make some data fs = 10; nsecs = 10; x1 = ao(plist('tsfcn', 'randn(size(t))', 'fs', fs, 'nsecs', nsecs, 'yunits', 'T')); x1.setName; x2 = ao(plist('tsfcn', 'randn(size(t))', 'fs', fs, 'nsecs', nsecs, 'yunits', 'm')); x2.setName; x3 = ao(plist('tsfcn', 'randn(size(t))', 'fs', fs, 'nsecs', nsecs, 'yunits', 'T')); x3.setName; x4 = ao(plist('tsfcn', 'randn(size(t))', 'fs', fs, 'nsecs', nsecs, 'yunits', 'm')); x4.setName; C1 = matrix(x1,x3,plist('shape',[2,1])); C1.setName; C2 = matrix(x2,x4,plist('shape',[2,1])); C2.setName; C = matrix([x1 x2;x3 x4]); C.setName; n1 = ao(plist('tsfcn', 'randn(size(t))', 'fs', fs, 'nsecs', nsecs, 'yunits', 'm')); n2 = ao(plist('tsfcn', 'randn(size(t))', 'fs', fs, 'nsecs', nsecs, 'yunits', 'm')); n = matrix(n1,n2,plist('shape',[2,1])); n.setName; a = [ao(1,plist('yunits','m/T')) ao(2,plist('yunits','m/m'))]; A = matrix(a,plist('shape',[2,1])); A.setName; y = C*A + n; %%% Get a fit with linlsqsvd pobj1 = linlsqsvd(C1, C2, y); % combine results for ii=1:numel(pobj1.y) prs(ii) = ao(cdata(pobj1.y(ii))); prs(ii).setYunits(pobj1.yunits(ii)); end Pars = matrix(prs,plist('shape',[numel(prs),1])); yfit1 = C*Pars; %%% do linear combination: using eval yfit2 = pobj1.eval; % Plot (compare data with fit) iplot(y.objs(1), yfit1.objs(1), yfit2.objs(1)) iplot(y.objs(2), yfit1.objs(2), yfit2.objs(2))