Testing TSCU by recreating the results in an article.

This script is prepared to test the Time Series Classification Utility by recreating the last column of Table II in the following article.

Gudmundsson, Steinn, Thomas Philip Runarsson, and Sven Sigurdsson.
"Support vector machines and dynamic time warping for time series."
Neural Networks, 2008. IJCNN 2008.(IEEE World Congress on Computational
Intelligence). IEEE International Joint Conference on. IEEE, 2008.
clear all
close all
clc

cSearch = 2.^(-5:2:15);
gSearch = 2.^(-15:2:3);
%         Dataset             error published in the article
data = { 'synthetic_control' , 2.33 ,...
         'Gun_Point'         ,12.67 ,...
         'CBF'               , 4.56 ,...
         'FaceAll'           ,26.57 ,...
         'OSULeaf'           ,40.08 ,...
         'SwedishLeaf'       ,38.24 ,...
         'Trace'             , 0.00 ,...
         'Two_Patterns'      , 0.00 ,...
         'wafer'             , 3.39 ,...
         'FaceFour'          ,11.36 ,...
         'Lighting2'         ,11.48 ,...
         'Lighting7'         ,30.14 ,...
         'ECG200'            ,17.00 ,...
         'Adiac'             ,56.01 ,...
         'yoga'              ,21.87 ,...
         'Fish'              ,29.71 ,...
         'Beef'              ,60.00 ,...
         'Coffee'            ,17.86 ,...
         'OliveOil'          ,26.67 ,...
};

fprintf('%-20s %-s\n'       ,''       ,'     SVM GDTW Error (%)');
fprintf('%-20s %-15s %-15s\n','Dataset','found by TSCU','in the Article');
fprintf('%-20s %-15s %-15s\n','-------','-------------','--------------');

for i=[1 2 18]
    dataname  = data{2*(i-1)+1};
    expected  = data{2*(i-1)+2};
    trn=load(sprintf('../../UCR/%s/%s_TRAIN',dataname,dataname));
    tst=load(sprintf('../../UCR/%s/%s_TEST',dataname,dataname));

    SVM_GDTW =tscu(trn,tst             ,...
        'Classifier'     ,'SVM'         ,...
        'Alignment'      ,'DTW'         ,...
        'SVMKernel'      ,'gaussian'    ,...
        'SVMSoftMargin'  ,cSearch       ,...
        'SVMGamma'       ,gSearch   ,...
        'CrossValidation',5             ,...
        'LogLevel'       ,'Alert');
    calculated = 100*(1-SVM_GDTW.perf.OA);
    fprintf('%-20s %-15.2f %-15.2f\n',dataname,calculated,expected);
end
                          SVM GDTW Error (%)
Dataset              found by TSCU   in the Article 
-------              -------------   -------------- 
synthetic_control    2.33            2.33           
Gun_Point            12.67           12.67          
Coffee               17.86           17.86