Predictor getTestPredictorLibSvm(){
//Load sample structure
SampleStructure sampleStructure = getTestSampleStructure();
//Load training data
Array<Sample> trainingData = getTrainingSampleData();
//Make creator lambdas
MlCreators creators;
creators.rc = new RegressorCreator[1];
creators.rc[0] = [](SampleStructure* st, Array<Sample> training, unsigned index){
svm_parameter svm_param;
svm_param.svm_type = EPSILON_SVR; // default for frac.libsvr is EPSILON_SVR, not C_SVC
svm_param.kernel_type = LINEAR; // noto changed default from RBF
svm_param.degree = 3;
svm_param.gamma = 0; // 1/num_features
svm_param.coef0 = 0;
svm_param.nu = 0.5;
svm_param.cache_size = 100;
svm_param.C = 1;
svm_param.eps = 1e-3;
svm_param.p = 0.0; // noto changed default from 0.1
svm_param.shrinking = 1;
svm_param.probability = 0;
svm_param.nr_weight = 0;
svm_param.weight_label = NULL;
svm_param.weight = NULL;
svm_param.timeout = 86400; // noto
Regressor* r = new SvmRegressor(*st, index, &svm_param);
r->train(training);
return r;
};
creators.cc = new ClassifierCreator[1];
creators.cc[0] = [](SampleStructure* st, Array<Sample> training, unsigned index){
Classifier* c = new WafflesDecisionTreeClassifier(*st, index, true);
c->train(training);
return c;
};
creators.bcc = new BinaryClassifierCreator[1];
creators.bcc[0] = [](SampleStructure* st, Array<Sample> training, unsigned index){
BinaryClassifier* r = new ConstantBinaryClassifier(false);
r->train(training);
return r;
};
//TODO build + use takeBest.
Predictor p = Predictor(sampleStructure, trainingData, creators);
return p;
}
int Process(ParametersType *Parameters, MatrixDataType *MatrixData, FemStructsType *FemStructs,
FemFunctionsType *FemFunctions, FemOtherFunctionsType *FemOtherFunctions)
{
int (*Predictor)(ParametersType *, MatrixDataType *, FemStructsType *, FemFunctionsType *, FemOtherFunctionsType *);
setProblem(Parameters, FemFunctions);
setMatrixVectorProductType(Parameters, FemFunctions);
setSolver(Parameters,FemOtherFunctions);
setPreconditioner(Parameters, FemFunctions);
setScaling(Parameters, FemFunctions);
setStabilizationForm(Parameters, FemFunctions, FemOtherFunctions, &Predictor);
setDimensionlessness(Parameters, FemFunctions);
set_BC_no_penetrability(Parameters, FemFunctions);
setStopCriteria(Parameters, FemFunctions);
Predictor(Parameters, MatrixData, FemStructs, FemFunctions, FemOtherFunctions);
return 0;
}
Predictor getTestPredictorWaffles(){
//Load sample structure
SampleStructure sampleStructure = getTestSampleStructure();
//Load training data
Array<Sample> trainingData = getTrainingSampleData();
//Make creator lambdas
MlCreators creators;
//Lambdas and side effects are best friends!
creators.rc = new RegressorCreator[1];
creators.rc[0] = [](SampleStructure* st, Array<Sample> training, unsigned index){
//Regressor* r = new ConstantRegressor(0);
Regressor* r = new WafflesDecisionTreeRegressor(*st, index);
r->train(training);
return r;
};
creators.cc = new ClassifierCreator[1];
creators.cc[0] = [](SampleStructure* st, Array<Sample> training, unsigned index){
Classifier* c = new WafflesDecisionTreeClassifier(*st, index, true); //TODO Why does this random give identical results to the default?
c->train(training);
return c;
};
creators.bcc = new BinaryClassifierCreator[1];
creators.bcc[0] = [](SampleStructure* st, Array<Sample> training, unsigned index){
BinaryClassifier* r = new ConstantBinaryClassifier(false);
r->train(training);
return r;
};
//TODO build + use takeBest.
//std::cout << "Making predictor." << std::endl;
Predictor p = Predictor(sampleStructure, trainingData, creators);
//std::cout << "Made predictor." << std::endl;
return p;
}
void ADERTimeStep(ADERDG* adg)
{
double dt = adg->dt_small;
// first, predict the values of w at an intermediate time step
for(int ie = 1;ie <= _NBELEMS_IN; ie++) {
Predictor(adg, ie, dt / 2);
}
// init the derivative to zero
for(int ie = 1; ie<= _NBELEMS_IN; ie++) {
for(int i = 0; i < _NGLOPS; i++) {
for(int iv = 0; iv < _M; iv++) {
adg->dtw[ie][i][iv]=0;
}
}
}
// impose the exact values on boundary left and right cells
double x = adg->face[0];
double t = adg->tnow + dt / 2 ;
ExactSol(x, t, adg->wpred[0][_D]); // _D = last point of left cell
x = adg->face[_NBFACES-1];
// 0 = first point of right cell
ExactSol(x, t, adg->wpred[_NBELEMS_IN + 1][0]);
// compute the face flux terms
// loop on the faces
for(int i = 0; i < _NBFACES; i++){
double *wL, *wR;
double flux[_M];
int ie = i;
wL = adg->wpred[ie][_D]; // _D = last point of left cell
wR = adg->wpred[ie+1][0]; // 0 = first point of right cell
NumFlux(wL, wR, flux);
for (int k = 0; k < _M; k++){
adg->dtw[ie][_D][k] -= flux[k];
adg->dtw[ie+1][0][k] += flux[k];
}
}
// compute the volume terms
for(int ie = 1; ie<= _NBELEMS_IN; ie++){
double h = adg->face[ie] - adg->face[ie-1];
// loop on the glops i
for(int i = 0; i < _NGLOPS; i++){
// integration weight
double omega = wglop(_D, i) * h;
// flux at glop i
double flux[_M];
NumFlux(adg->wpred[ie][i], adg->wpred[ie][i], flux);
// loop on the basis functions j
for(int j = 0; j < _D+1; j++){
// derivative of basis function j at glop i
double dd = dlag(_D, j, i) / h;
for (int k = 0; k < _M; k++){
adg->dtw[ie][j][k] += omega * dd * flux[k];
}
}
}
}
// divide by the mass matrix
for(int ie = 1; ie<= _NBELEMS_IN; ie++){
double h = adg->face[ie] - adg->face[ie-1];
for(int i = 0; i < _NGLOPS; i++){
double omega = wglop(_D, i) * h;
for (int k = 0; k < _M; k++){
adg->dtw[ie][i][k] /= omega;
}
}
}
// update wnext and
// copy wnext into wnow for the next time step
for(int ie = 1; ie<= _NBELEMS_IN; ie++){
for(int i = 0; i < _NGLOPS; i++){
for(int iv = 0; iv < _M; iv++){
adg->wnext[ie][i][iv] =
adg->wnow[ie][i][iv] + dt * adg->dtw[ie][i][iv];
adg->wnow[ie][i][iv] = adg->wnext[ie][i][iv];
}
}
}
}
