FrequencyLearning.m

Go to the documentation of this file.

namespace models{
namespace motoneuron{
namespace experiments{


/* (Autoinserted by mtoc++)
 * This source code has been filtered by the mtoc++ executable,
 * which generates code that can be processed by the doxygen documentation tool.
 *
 * On the other hand, it can neither be interpreted by MATLAB, nor can it be compiled with a C++ compiler.
 * Except for the comments, the function bodies of your M-file functions are untouched.
 * Consequently, the FILTER_SOURCE_FILES doxygen switch (default in our Doxyfile.template) will produce
 * attached source files that are highly readable by humans.
 *
 * Additionally, links in the doxygen generated documentation to the source code of functions and class members refer to
 * the correct locations in the source code browser.
 * However, the line numbers most likely do not correspond to the line numbers in the original MATLAB source files.
 */

function  FrequencyLearning() {
clear classes;
basedir = fileparts(mfilename(" fullpath "));

file = fullfile(basedir, " paramdomaindetection_withnoise.mat ");
load(file);

/* % Plot raw data */
pm = PlotManager;
pm.FigureSize= [1024 768];
ax = pm.nextPlot(" raw_data "," Frequencies by detector "," Fibre type \tau "," Mean current \kappa ");
tri = delaunay(ps(1,:),ps(2,:));
trisurf(tri,ps(1,:),ps(2,:),Hz," Parent ",ax," FaceColor "," interp "," EdgeColor "," interp ");

/* % assemble training data */
m = min(ps,[],2);
M = max(ps,[],2);
n = 200;
ft = linspace(m(1),M(1),n);
mc = linspace(m(2),M(2),n);
[FT,MC] = meshgrid(ft,mc);
HZ = griddata(ps(1,:),ps(2,:),Hz,FT,MC," nearest ");

/* % smooth out data */
padsize = 6;
stencil = [0 0 1 0 0; 0 1 1 1 0; 1 1 1 1 1; 0 1 1 1 0; 0 0 1 0 0];
/*  stencil = [1 1 1; 1 1 1; 1 1 1];
 * stencil = [0 1 0; 1 1 1; 0 1 0]; */
stencil = stencil/sum(stencil(:));
HZ = padarray(HZ,[padsize padsize]," replicate ");
runs = 10;
for k = 1:runs
    HZ = conv2(HZ,stencil," same ");
end
HZ = HZ(padsize+1:end-padsize,padsize+1:end-padsize);
ax2 = pm.nextPlot(" smoothed "," Smoothed frequency surface "," Fibre type \tau "," Mean current \kappa ");
surf(ax2,FT,MC,HZ," FaceColor "," interp "," EdgeColor "," interp ");
axis(ax2," tight ");

/* % Init */
s = load(fullfile(basedir," ../ "," upperlimitpoly.mat "));
maxv = polyval(s.upperlimit_poly,FT);
valid = MC <= maxv;
atd = data.ApproxTrainData([FT(valid) MC(valid)]^t,[],[]);
atd.fxi= HZ(valid)^t;

/* % init algorithm */
alg = approx.algorithms.VKOGA;
ec = kernels.config.ExpansionConfig;
comb = Utils.createCombinations([1 2 3 4],[1 2 3]);
sc = kernels.config.WendlandConfig(" G ",comb(1,:)," S ",comb(2,:));
ec.StateConfig= sc;
alg.ExpConfig= ec;
alg.MaxExpansionSize= 1000;

/* % compute */
kexp = alg.computeApproximation(atd);

/* % plot */
afx = kexp.evaluate([FT(:) MC(:)]^t);
FX = reshape(afx,200,[]);
FX(~valid) = Inf;
ax3 = pm.nextPlot(" approx "," Learned smoothed frequency surface "," Fibre type \tau "," Mean current \kappa ");
surf(ax3,FT,MC,FX," FaceColor "," interp "," EdgeColor "," interp ");
axis(ax3," tight ");

/* % Save stuff */
pm.UseFileTypeFolders= false;
pm.ExportDPI= 150;
pm.FilePrefix= " freq ";
pm.NoTitlesOnSave= true;
zlim(ax2,zlim(ax));
zlim(ax3,zlim(ax));
pm.savePlots(basedir," Format "," jpg "," Close ",true);
file = fullfile(basedir, " FreqencyLearning.mat ");
save(file, " alg ", " atd ", " HZ ", " FT ", " MC ", " valid ");
file = fullfile(basedir, " FrequencyKexp.mat ");
save(file, " kexp ");
}

};
};
};