function RunSVMClassifierSINK(DataSetStartIndex, DataSetEndIndex)

% first 2 values are '.' and '..' - UCR Archive 2018 version has 128 datasets

dir_struct = dir('/rigel/dsi/users/ikp2103/VLDBGRAIL/UCR2018/');

Datasets = {dir_struct(3:130).name};

% Sort Datasets

[Datasets, DSOrder] = sort(Datasets);

addpath(genpath('SVMMatlab/.'));

distcomp.feature( 'LocalUseMpiexec', false )

rng(ceil(DataSetStartIndex*100))

pause(300*rand);

poolobj = gcp('nocreate');

delete(poolobj);

parpool(22);

rng('default')

for i = 1:length(Datasets)

if (i>=DataSetStartIndex && i<=DataSetEndIndex)

Results = zeros(length(Datasets),14);

disp(['Dataset being processed: ', char(Datasets(i))]);

DS = LoadUCRdataset(char(Datasets(i)));

%TrainInstancesCount = 1000;

%TrainClassLabels = DS.TrainClassLabels(1:1000);

TrainInstancesCount = DS.TrainInstancesCount;

TrainClassLabels = DS.TrainClassLabels;

Thebestgamma1FINAL = 0;

Thebestcost1FINAL = 0;

Thebestacc1FINAL = 0;

Thebestiming1FINAL = 0;

ThebestWarpFINAL = 0;

% optimize additional parameters. For now turned off

for WarpValue = 0:0

[Thebestgamma1,Thebestcost1,Thebestacc1,Thebestiming1] = GridSearchSVM1(-10,0.1,20,TrainInstancesCount,TrainClassLabels,Datasets,i,WarpValue);

if Thebestacc1>Thebestacc1FINAL

Thebestacc1FINAL = Thebestacc1;

Thebestgamma1FINAL = Thebestgamma1;

Thebestcost1FINAL = Thebestcost1;

ThebestWarpFINAL = WarpValue;

Thebestiming1FINAL = Thebestiming1FINAL Thebestiming1;

end

end

[Thebestgamma2,Thebestcost2,Thebestacc2,Thebestiming2] = GridSearchSVM2(Thebestcost1FINAL-2,0.01,Thebestcost1FINAL 2,TrainInstancesCount,TrainClassLabels,Datasets,i,Thebestgamma1FINAL,ThebestWarpFINAL);

[Thebestgamma3,Thebestcost3,Thebestacc3,Thebestiming3] = GridSearchSVM2(Thebestcost2-0.2,0.001,Thebestcost2 0.2,TrainInstancesCount,TrainClassLabels,Datasets,i,Thebestgamma1FINAL,ThebestWarpFINAL);

KM = dlmread( strcat( 'KernelMatricesSINK/',char(Datasets(i)),'/', char(Datasets(i)), '_SINK_Gamma_', num2str(Thebestgamma3),'.kernelmatrix') );

KMTrain = KM(1:DS.TrainInstancesCount,1:DS.TrainInstancesCount);

tic;

cmd = ['-q -m 500 -t 4 -e 0.001 -c ', num2str(2^Thebestcost3)];

model_precomputed = svmtrain(DS.TrainClassLabels, [(1:DS.TrainInstancesCount)', KMTrain], cmd);

ModelTrainingRuntime = toc;

KM = dlmread( strcat( 'KernelMatricesSINK/',char(Datasets(i)),'/', char(Datasets(i)), '_SINK_Gamma_', num2str(Thebestgamma3),'.kernelmatrix') );

KMTESTTOTRAIN = KM(DS.TrainInstancesCount 1:end,1:DS.TrainInstancesCount);

tic;

%[predict_label_P, accuracy_P, dec_values_P] = svmpredict(DS.TestClassLabels, [(1:DS.TestInstancesCount)', KM(DS.TrainInstancesCount 1:end,1:DS.TrainInstancesCount)], model_precomputed);

[predict_label_P, accuracy_P, dec_values_P] = svmpredict(DS.TestClassLabels, [(1:DS.TestInstancesCount)', KMTESTTOTRAIN], model_precomputed);

%[predict_label_P, accuracy_P, dec_values_P] = svmpredict(DS.TestClassLabels(1:1000), [(1:1000)', KMTESTTOTRAIN], model_precomputed);

PredictionRuntime = toc;

Results(i,1) = ThebestWarpFINAL;

Results(i,2) = Thebestgamma1FINAL;

Results(i,3) = Thebestgamma2;

Results(i,4) = Thebestgamma3;

Results(i,5) = Thebestcost1FINAL;

Results(i,6) = Thebestcost2;

Results(i,7) = Thebestcost3;

Results(i,8) = Thebestacc1FINAL*0.01;

Results(i,9) = Thebestacc2*0.01;

Results(i,10) = Thebestacc3*0.01;

Results(i,11) = Thebestiming1FINAL Thebestiming2 Thebestiming3;

Results(i,12) = accuracy_P(1)*0.01;

Results(i,13) = ModelTrainingRuntime;

Results(i,14) = PredictionRuntime;

dlmwrite( strcat('RunSVMClassifierSINK/','RunSVMClassifierSINK_', num2str(i)), Results, 'delimiter', '\t');

end

end

poolobj = gcp('nocreate');

delete(poolobj);

end

function [Thebestgamma,Thebestcost,Thebestacc,Thebestiming] = GridSearchSVM1(GridStart,GridStep,GridEnd,TrainInstancesCount,TrainClassLabels,Datasets,DatasetsNumber,WarpValue)

previousMaxbestacc = 0;

Thebestgamma = 0;

Thebestcost = 0;

Thebestacc = 0;

Thebestiming = 0;

% Tuning Parameters

for gamma=1:20

gamma

log2cTmp = GridStart:GridStep:GridEnd;

bestacc = zeros(1,length(log2cTmp));

bestgamma = zeros(1,length(log2cTmp));

bestcost = zeros(1,length(log2cTmp));

besttiming = zeros(1,length(log2cTmp));

KM = dlmread( strcat( 'KernelMatricesSINK/',char(Datasets(DatasetsNumber)),'/', char(Datasets(DatasetsNumber)), '_SINK_Gamma_', num2str(gamma),'.kernelmatrix') );

KMTrain = KM(1:TrainInstancesCount,1:TrainInstancesCount);

% grid search

parfor log2cNEW = 1:length(log2cTmp)

log2cNEW

tic;

log2c = log2cTmp(log2cNEW);

cmd = ['-q -m 500 -t 4 -e 0.001 -v ' num2str(10) ' -c ', num2str(2^log2c)];

cv = svmtrain(TrainClassLabels, [(1:TrainInstancesCount)', KMTrain], cmd);

bestacc(log2cNEW) = cv;

bestcost(log2cNEW) = log2c;

bestgamma(log2cNEW) = gamma;

besttiming(log2cNEW) = toc;

end

[Maxbestacc,~] = max(bestacc);

Posbestacc = find(bestacc==Maxbestacc,1,'last');

Thebestiming = Thebestiming sum(besttiming);

if Maxbestacc>previousMaxbestacc

Thebestgamma = bestgamma(Posbestacc);

Thebestcost = bestcost(Posbestacc);

Thebestacc = Maxbestacc;

previousMaxbestacc = Maxbestacc;

end

end

end

function [Thebestgamma,Thebestcost,Thebestacc,Thebestiming] = GridSearchSVM2(GridStart,GridStep,GridEnd,TrainInstancesCount,TrainClassLabels,Datasets,DatasetsNumber,gamma,WarpValue)

previousMaxbestacc = 0;

Thebestgamma = 0;

Thebestcost = 0;

Thebestacc = 0;

Thebestiming = 0;

% Tuning Parameters

log2cTmp = GridStart:GridStep:GridEnd;

bestacc = zeros(1,length(log2cTmp));

bestgamma = zeros(1,length(log2cTmp));

bestcost = zeros(1,length(log2cTmp));

besttiming = zeros(1,length(log2cTmp));

KM = dlmread( strcat( 'KernelMatricesSINK/',char(Datasets(DatasetsNumber)),'/', char(Datasets(DatasetsNumber)), '_SINK_Gamma_', num2str(gamma),'.kernelmatrix') );

KMTrain = KM(1:TrainInstancesCount,1:TrainInstancesCount);

% grid search

parfor log2cNEW = 1:length(log2cTmp)

log2cNEW

tic;

log2c = log2cTmp(log2cNEW);

cmd = ['-q -m 500 -t 4 -e 0.001 -v ' num2str(10) ' -c ', num2str(2^log2c)];

cv = svmtrain(TrainClassLabels, [(1:TrainInstancesCount)', KMTrain], cmd);

bestacc(log2cNEW) = cv;

bestcost(log2cNEW) = log2c;

bestgamma(log2cNEW) = gamma;

besttiming(log2cNEW) = toc;

end

[Maxbestacc,~] = max(bestacc);

Posbestacc = find(bestacc==Maxbestacc,1,'last');

Thebestiming = Thebestiming sum(besttiming);

if Maxbestacc>previousMaxbestacc

Thebestgamma = bestgamma(Posbestacc);

Thebestcost = bestcost(Posbestacc);

Thebestacc = Maxbestacc;

previousMaxbestacc = Maxbestacc;

end

end