CFactorGraph CompressInterface::createCFactorGraph() {
createVarClustering();
createFacClustering();
// create lifted fg here
vector<CFactor> superFacs;
// clusterIdx => facIdx
for (map<size_t,size_t>::iterator facIter = _facRepr.begin(); facIter != _facRepr.end(); facIter++) {
VarSet superVarSet;
foreach (const dai::BipartiteGraph::Neighbor &tmpVar, _cfg.nbF(facIter->second)) {
Var varCluster = _varRepr[_varColorVec[tmpVar]];
if (!superVarSet.contains(varCluster)) {
superVarSet |= Var(varCluster);
}
}
CFactor superFac = CFactor(superVarSet, _cfg.factor(facIter->second).p());
superFac.sigma() = _cfg.factor(facIter->second).sigma();
superFac.position() = _cfg.factor(facIter->second).position();
superFac.counts() = createCounts(facIter->second, superVarSet);
superFacs.push_back(superFac);
}
return CFactorGraph(superFacs);
}
void CLWSamplingInfEngine::
EnterEvidence( const CEvidence *pEvidenceIn , int maximize , int sumOnMixtureNode )
{
PNL_CHECK_IS_NULL_POINTER(pEvidenceIn);
LWSampling(pEvidenceIn);
// Given evidencs, calculate particle weight by CPD
float w;
int iSampleSize = m_currentEvVec.size();
const int* ObsNodes = pEvidenceIn->GetAllObsNodes();
int NumberObsNodes = pEvidenceIn->GetNumberObsNodes();
int i, j;
for( i = 0; i < iSampleSize; i++)
{
w = 0;
for( j = 0; j < NumberObsNodes; j++)
{
if(pEvidenceIn->IsNodeObserved(ObsNodes[j]))
{
CFactor* pFactor = m_pGraphicalModel->GetFactor(ObsNodes[j]);
w = w + pFactor->GetLogLik( m_currentEvVec[i]);
}
}
m_particleWeight[i] = (float)exp(w);
}
NormalizeWeight();
}
void CFactorGraph::compute_energies()
{
for (int32_t fi = 0; fi < m_factors->get_num_elements(); ++fi)
{
CFactor* fac = dynamic_cast<CFactor*>(m_factors->get_element(fi));
fac->compute_energies();
SG_UNREF(fac);
}
}
void CFactorGraph::loss_augmentation(SGVector<int32_t> states_gt, SGVector<float64_t> loss)
{
if (loss.size() == 0)
{
loss.resize_vector(states_gt.size());
SGVector<float64_t>::fill_vector(loss.vector, loss.vlen, 1.0 / states_gt.size());
}
int32_t num_vars = states_gt.size();
ASSERT(num_vars == loss.size());
SGVector<int32_t> var_flags(num_vars);
var_flags.zero();
// augment loss to incorrect states in the first factor containing the variable
// since += L_i for each variable if it takes wrong state ever
// TODO: augment unary factors
for (int32_t fi = 0; fi < m_factors->get_num_elements(); ++fi)
{
CFactor* fac = dynamic_cast<CFactor*>(m_factors->get_element(fi));
SGVector<int32_t> vars = fac->get_variables();
for (int32_t vi = 0; vi < vars.size(); vi++)
{
int32_t vv = vars[vi];
ASSERT(vv < num_vars);
if (var_flags[vv])
continue;
SGVector<float64_t> energies = fac->get_energies();
for (int32_t ei = 0; ei < energies.size(); ei++)
{
CTableFactorType* ftype = fac->get_factor_type();
int32_t vstate = ftype->state_from_index(ei, vi);
SG_UNREF(ftype);
if (states_gt[vv] == vstate)
continue;
// -delta(y_n, y_i_n)
fac->set_energy(ei, energies[ei] - loss[vv]);
}
var_flags[vv] = 1;
}
SG_UNREF(fac);
}
// make sure all variables have been checked
int32_t min_var = SGVector<int32_t>::min(var_flags.vector, var_flags.vlen);
ASSERT(min_var == 1);
}
float64_t CFactorGraph::evaluate_energy(const SGVector<int32_t> state) const
{
ASSERT(state.size() == m_cardinalities.size());
float64_t energy = 0.0;
for (int32_t fi = 0; fi < m_factors->get_num_elements(); ++fi)
{
CFactor* fac = dynamic_cast<CFactor*>(m_factors->get_element(fi));
energy += fac->evaluate_energy(state);
SG_UNREF(fac);
}
return energy;
}
void CompressInterface::rearrangeCnfFactor(CFactor &cnfClause) {
State state(cnfClause.vars());
vector<size_t> (cnfClause.vars().size());
vector<size_t> pos,neg,sigma;
for (size_t i=0; i<cnfClause.states(); i++) {
if (cnfClause[i] == 1) {
for (vector<Var>::const_iterator varIter=cnfClause.vars().begin(); varIter!=cnfClause.vars().end();varIter++) {
if (state(*varIter) == 0) {
pos.push_back(varIter - cnfClause.vars().begin());
} else {
neg.push_back(varIter - cnfClause.vars().begin());
}
}
break;
} else if (cnfClause[i] == 0) {
//TODO handle factors which contain zero states
// in such cases the possible zero state could have been clamped;
// possible solutions:
// i) restore factor from backup (requires to clamp factors accordingly)
// ii) cache zero states in advance
return;
}
state++;
}
sigma.insert(sigma.begin(),pos.begin(), pos.end());
sigma.insert(sigma.begin() + pos.size(),neg.begin(), neg.end());
cnfClause.sigma() = sigma;
}
void CBayesLearningEngine::Learn()
{
if(!m_pGrModel)
{
PNL_THROW( CNULLPointer, "no graphical model")
}
CStaticGraphicalModel *grmodel = this->GetStaticModel();
CFactor *factor = NULL;
int numberOfFactors = grmodel->GetNumberOfFactors();
int domainNodes;
if(m_numberOfLearnedEvidences == m_numberOfAllEvidences)
{
PNL_THROW(COutOfRange, "number of unlearned evidences must be positive")
}
int currentEvidNumber;
const CEvidence* pCurrentEvid;
//below code is intended to work on tabular CPD and gaussian CPD
//later we will generalize it for other distribution types
if ((grmodel->GetFactor(0))->GetDistributionType() == dtTabular)
{
for( int ev = m_numberOfLearnedEvidences; ev < m_numberOfAllEvidences; ev++)
{
currentEvidNumber = ev;
pCurrentEvid = m_Vector_pEvidences[currentEvidNumber];
if( !pCurrentEvid)
{
PNL_THROW(CNULLPointer, "evidence")
}
for( domainNodes = 0; domainNodes < numberOfFactors; domainNodes++ )
{
factor = grmodel->GetFactor( domainNodes );
int DomainSize;
const int *domain;
factor->GetDomain( &DomainSize, &domain );
const CEvidence *pEvidences[] = { pCurrentEvid };
CTabularDistribFun* pDistribFun = (CTabularDistribFun*)(factor->GetDistribFun());
pDistribFun->BayesUpdateFactor(pEvidences, 1, domain);
}
}
}
else
{
for( domainNodes = 0; domainNodes < numberOfFactors; domainNodes++ )
SGVector< float64_t > CFactorGraphModel::get_joint_feature_vector(int32_t feat_idx, CStructuredData* y)
{
// factor graph instance
CFactorGraphFeatures* mf = CFactorGraphFeatures::obtain_from_generic(m_features);
CFactorGraph* fg = mf->get_sample(feat_idx);
// ground truth states
CFactorGraphObservation* fg_states = CFactorGraphObservation::obtain_from_generic(y);
SGVector<int32_t> states = fg_states->get_data();
// initialize psi
SGVector<float64_t> psi(get_dim());
psi.zero();
// construct unnormalized psi
CDynamicObjectArray* facs = fg->get_factors();
for (int32_t fi = 0; fi < facs->get_num_elements(); ++fi)
{
CFactor* fac = dynamic_cast<CFactor*>(facs->get_element(fi));
CTableFactorType* ftype = fac->get_factor_type();
int32_t id = ftype->get_type_id();
SGVector<int32_t> w_map = get_params_mapping(id);
ASSERT(w_map.size() == ftype->get_w_dim());
SGVector<float64_t> dat = fac->get_data();
int32_t dat_size = dat.size();
ASSERT(w_map.size() == dat_size * ftype->get_num_assignments());
int32_t ei = ftype->index_from_universe_assignment(states, fac->get_variables());
for (int32_t di = 0; di < dat_size; di++)
psi[w_map[ei*dat_size + di]] += dat[di];
SG_UNREF(ftype);
SG_UNREF(fac);
}
// negation (-E(x,y) = <w,phi(x,y)>)
psi.scale(-1.0);
SG_UNREF(facs);
SG_UNREF(fg);
return psi;
}
CMNet* CMNet::CreateWithRandomMatrices( const intVecVector& clqs,
CModelDomain* pMD)
{
CMNet* pMNet = CMNet::Create( clqs, pMD );
pMNet->AllocFactors();
int numFactors = pMNet->GetNumberOfFactors();
int i;
for( i = 0; i < numFactors; i++ )
{
pMNet->AllocFactor( i );
CFactor* ft = pMNet->GetFactor(i);
ft->CreateAllNecessaryMatrices();
}
return pMNet;
}
void cfactor_setitem(CFactor& factor, size_t index, double value)
{
if (index >= 0 && index < factor.states()) {
factor[index] = value;
}
else {
PyErr_SetString(PyExc_IndexError, "index out of range");
boost::python::throw_error_already_set();
}
}
//create mnet with random matrices
CMRF2* CMRF2::CreateWithRandomMatrices( int numberOfCliques,
const int *cliqueSizes,
const int **cliques,
CModelDomain* pMD)
{
CMRF2* pMRF2 = CMRF2::Create( numberOfCliques, cliqueSizes, cliques, pMD );
pMRF2->AllocFactors();
int numFactors = pMRF2->GetNumberOfFactors();
int i;
for( i = 0; i < numFactors; i++ )
{
pMRF2->AllocFactor( i );
CFactor* ft = pMRF2->GetFactor(i);
ft->CreateAllNecessaryMatrices();
}
return pMRF2;
}
void CMlLearningEngine::Learn()
{
/*
function takes an information from m_pEvidences and learns factors
of graphical model using prior probabilities or not
*/
float logLikTmp = 0;
if(!m_pGrModel)
{
PNL_THROW( CNULLPointer, "no graphical model")
}
CStaticGraphicalModel *grmodel = this->GetStaticModel();
CFactor *parameter = NULL;
int numberOfDomains = grmodel -> GetNumberOfFactors();
for( int domainNodes = 0; domainNodes < numberOfDomains; domainNodes++ )
{
factor = grmodel->GetFactor( domainNodes );
factor ->UpdateStatisticsML( &m_Vector_pEvidences.front(),
m_Vector_pEvidences.size() );
PNL_CHECK_LEFT_BORDER(m_numberOfAllEvidences, 1);
logLikTmp += parameter->ProcessingStatisticalData(m_numberOfAllEvidences);
}
switch( grmodel -> GetModelType() )
{
case mtBNet:
{
break;
}
case mtMRF2:
case mtMNet:
{
logLikTmp = _LearnPotentials();
break;
}
default:
{
PNL_THROW(CBadConst, "model type" )
break;
}
}
m_critValue.push_back(logLikTmp);
}
double cfactor_getitem(CFactor &factor, size_t index)
{
if (index >= 0 && index < factor.states()) {
return factor[index];
}
else {
PyErr_SetString(PyExc_IndexError, "index out of range");
boost::python::throw_error_already_set();
return -1;
}
}
CCPD* CStaticStructLearnSEM::CreateRandomCPD(int nfamily, const int* family, CBNet* pBNet)
{
int child = family[nfamily-1];
CModelDomain* pMD = pBNet->GetModelDomain();
CFactor* factor = pBNet->GetFactor(child);
EDistributionType dt = factor->GetDistributionType();
CCPD* pCPD;
if( dt == dtTabular )
{
pCPD = CTabularCPD::Create(family, nfamily, pMD);
pCPD->CreateAllNecessaryMatrices(1);
return pCPD;
}
else
{
if( dt == dtMixGaussian )
{
floatVector data;
static_cast<CMixtureGaussianCPD*>(factor)->GetProbabilities(&data);
pCPD = CMixtureGaussianCPD::Create(family, nfamily, pMD, &data.front());
static_cast<CCondGaussianDistribFun*>(pCPD->GetDistribFun()) -> CreateDefaultMatrices(1);
return pCPD;
}
else
{
if( (dt == dtGaussian) || (dt == dtCondGaussian) )
{
pCPD = CGaussianCPD::Create(family, nfamily, pMD);
pCPD->CreateAllNecessaryMatrices(1);
return pCPD;
}
else
PNL_THROW(CNotImplemented, "this type of distribution is not supported yet");
}
}
}
void CFactorGraph::connect_components()
{
if (m_dset->get_connected())
return;
// need to be reset once factor graph is updated
m_dset->make_sets();
bool flag = false;
for (int32_t fi = 0; fi < m_factors->get_num_elements(); ++fi)
{
CFactor* fac = dynamic_cast<CFactor*>(m_factors->get_element(fi));
SGVector<int32_t> vars = fac->get_variables();
int32_t r0 = m_dset->find_set(vars[0]);
for (int32_t vi = 1; vi < vars.size(); vi++)
{
// for two nodes in a factor, should be an edge between them
// but this time link() isn't performed, if they are linked already
// means there is another path connected them, so cycle detected
int32_t ri = m_dset->find_set(vars[vi]);
if (r0 == ri)
{
flag = true;
continue;
}
r0 = m_dset->link_set(r0, ri);
}
SG_UNREF(fac);
}
m_has_cycle = flag;
m_dset->set_connected(true);
}
void CParEMLearningEngine::LearnOMP()
{
CStaticGraphicalModel *pGrModel = this->GetStaticModel();
PNL_CHECK_IS_NULL_POINTER(pGrModel);
PNL_CHECK_LEFT_BORDER(GetNumEv() - GetNumberProcEv() , 1);
//omp_set_num_threads(2);
int numberOfThreads = omp_get_num_procs();
//CParPearlInfEngine **pCurrentInfEng = new CParPearlInfEngine*[numberOfThreads];
CJtreeInfEngine **pCurrentInfEng = new CJtreeInfEngine*[numberOfThreads];
for (int i = 0; i < numberOfThreads; i++)
pCurrentInfEng[i] = NULL;
CFactor *parameter1 = NULL;
int exit = 0;
int numberOfParameters = pGrModel->GetNumberOfParameters();
int domainNodes;
//int itsML = 0;
// !!!
float loglik = -FLT_MAX;
float loglikOld = -FLT_MAX;
float epsilon = GetPrecisionEM();
float stopExpression = epsilon + 1.0f;
int iteration = 0;
int ev;
// to create additional factors
CFactor **ppAllFactors = new CFactor*[numberOfParameters*numberOfThreads];
bool *was_updated = new bool[numberOfParameters*numberOfThreads];
int factor;
#pragma omp parallel for private(factor) default(shared)
for (factor = 0; factor < numberOfParameters; factor++)
{
ppAllFactors[factor] = pGrModel->GetFactor(factor);
ppAllFactors[factor]->GetDistribFun()->ClearStatisticalData();
was_updated[factor] = false;
for (int proc = 1; proc < numberOfThreads; proc++)
{
ppAllFactors[factor + proc * numberOfParameters] =
ppAllFactors[factor]->Clone();
ppAllFactors[factor + proc * numberOfParameters]->GetDistribFun()->
ClearStatisticalData();
was_updated[factor + proc * numberOfParameters]= false;
};
};
int* itsML = new int[numberOfThreads];
for (int delta = 0; delta < numberOfThreads; delta++)
{
itsML[delta] = 0;
};
int start_ev, end_ev;
do
{
iteration++;
start_ev = GetNumberProcEv();
end_ev = GetNumEv();
#pragma omp parallel for schedule(dynamic) private(ev)
for (ev = start_ev; ev < end_ev ; ev++)
{
CFactor *parameter = NULL;
int DomainNodes_new;
int bMaximize = 0;
int bSumOnMixtureNode = 0;
int infIsNeed = 0;
int currentEvidNumber = ev; // !!!
const CEvidence* pCurrentEvid = m_Vector_pEvidences[currentEvidNumber];
infIsNeed = !GetObsFlags(ev)->empty(); // !!!
int Num_thread = omp_get_thread_num();
if (infIsNeed)
{
if (!pCurrentInfEng[Num_thread])
{
pCurrentInfEng[Num_thread] = CJtreeInfEngine::Create(
(const CStaticGraphicalModel *)pGrModel);
}
pCurrentInfEng[Num_thread]->EnterEvidence(pCurrentEvid, bMaximize,
bSumOnMixtureNode);
}
for (DomainNodes_new = 0; DomainNodes_new < numberOfParameters;
DomainNodes_new++)
{
parameter = ppAllFactors[DomainNodes_new +
Num_thread * numberOfParameters];
if (infIsNeed)
{
int DomainSize;
const int *domain;
parameter->GetDomain(&DomainSize, &domain);
if (IsDomainObserved(DomainSize, domain, currentEvidNumber))
{
const CEvidence *pEvidences[] = { pCurrentEvid };
parameter->UpdateStatisticsML(pEvidences, 1);
was_updated[DomainNodes_new+Num_thread*numberOfParameters]= true;
}
else
{
pCurrentInfEng[Num_thread]->MarginalNodes(domain, DomainSize, 1);
const CPotential * pMargPot =
pCurrentInfEng[Num_thread]->GetQueryJPD();
parameter ->UpdateStatisticsEM(pMargPot, pCurrentEvid);
was_updated[DomainNodes_new+Num_thread*numberOfParameters]= true;
}
}
else
{
const CEvidence *pEvidences[] = { pCurrentEvid };
parameter->UpdateStatisticsML(pEvidences, 1);
was_updated[DomainNodes_new+Num_thread*numberOfParameters]= true;
}
}
itsML[Num_thread] = itsML[Num_thread] || !infIsNeed;
} // end of parallel for
for (int delta = 1; delta < numberOfThreads; delta++)
{
itsML[0] = itsML[0] || itsML[delta];
};
//to join factors
#pragma omp parallel for private(factor) default(shared)
for (factor = 0; factor < numberOfParameters; factor++)
{
for (int proc = 1; proc < numberOfThreads; proc++)
{
if (was_updated[factor+proc*numberOfParameters])
{
ppAllFactors[factor]->UpdateStatisticsML(ppAllFactors[factor +
proc*numberOfParameters]);
ppAllFactors[factor+proc*numberOfParameters]->GetDistribFun()->
ClearStatisticalData();
};
was_updated[factor+proc*numberOfParameters] = false;
};
};
switch (pGrModel->GetModelType())
{
case mtBNet:
{
loglikOld = loglik;
loglik = 0.0f;
for (domainNodes = 0; domainNodes < numberOfParameters; domainNodes++)
{
parameter1 = pGrModel->GetFactor(domainNodes);
loglik += parameter1->ProcessingStatisticalData(
m_numberOfAllEvidences);
}
break;
}
case mtMRF2:
case mtMNet:
{
loglikOld = loglik;
loglik = _LearnPotentials();
break;
}
default:
{
PNL_THROW(CBadConst, "model type")
break;
}
}
stopExpression = float(fabs(2 * (loglikOld - loglik) /
(loglikOld + loglik)));
exit = ((stopExpression > epsilon) && (iteration <= GetMaxIterEM())) && !itsML[0];
if (exit)
{
ClearStatisticData();
}
m_critValue.push_back(loglik);
for (int j = 0; j < numberOfThreads; j++)
{
delete pCurrentInfEng[j];
pCurrentInfEng[j] = NULL;
}
} while (exit);
delete [] pCurrentInfEng;
//”даление дополнительных факторов
for (factor = numberOfParameters; factor < numberOfParameters * numberOfThreads;
factor++)
{
delete ppAllFactors[factor];
};
delete[] ppAllFactors;
delete[] was_updated;
if (iteration > GetMaxIterEM())
{
PNL_THROW(CNotConverged, "maximum number of iterations")
}
SetNumProcEv( GetNumEv() );
}
void CParEMLearningEngine::Learn()
{
CStaticGraphicalModel *pGrModel = this->GetStaticModel();
PNL_CHECK_IS_NULL_POINTER(pGrModel);
PNL_CHECK_LEFT_BORDER(GetNumEv() - GetNumberProcEv() , 1);
CJtreeInfEngine *pCurrentInfEng = NULL;
CFactor *parameter = NULL;
int exit = 0;
int numberOfParameters = pGrModel->GetNumberOfParameters();
int domainNodes;
int infIsNeed = 0;
int itsML = 0;
// !!!
float loglik = -FLT_MAX;
float loglikOld = -FLT_MAX;
float epsilon = GetPrecisionEM();
float stopExpression = epsilon + 1.0f;
int iteration = 0;
int currentEvidNumber;
int bMaximize = 0;
int bSumOnMixtureNode = 0;
const CEvidence* pCurrentEvid;
int start_mpi, finish_mpi;
int NumberOfProcesses, MyRank;
int numSelfEvidences;
MPI_Comm_size(MPI_COMM_WORLD, &NumberOfProcesses);
MPI_Comm_rank(MPI_COMM_WORLD, &MyRank);
int d = 0;
do
{
iteration++;
numSelfEvidences = (GetNumEv() - GetNumberProcEv()) / NumberOfProcesses;
start_mpi = GetNumberProcEv() + numSelfEvidences * MyRank; // !!!
if (MyRank < NumberOfProcesses - 1)
finish_mpi = start_mpi + numSelfEvidences; // !!!
else
finish_mpi = GetNumEv(); // !!!
for(int ev = start_mpi; ev < finish_mpi; ev++)
{
infIsNeed = 0;
currentEvidNumber = ev; // !!!
pCurrentEvid = m_Vector_pEvidences[currentEvidNumber];
if( !pCurrentEvid)
{
PNL_THROW(CNULLPointer, "evidence")
}
infIsNeed = !GetObsFlags(ev)->empty(); // !!!
if(infIsNeed)
{
// create inference engine
if(!pCurrentInfEng)
{
pCurrentInfEng = CJtreeInfEngine::Create(pGrModel);
}
pCurrentInfEng->EnterEvidence(pCurrentEvid, bMaximize,
bSumOnMixtureNode);
}
for(domainNodes = 0; domainNodes < numberOfParameters; domainNodes++)
{
parameter = pGrModel->GetFactor(domainNodes);
if(infIsNeed)
{
int DomainSize;
const int *domain;
parameter->GetDomain(&DomainSize, &domain);
if (IsDomainObserved(DomainSize, domain, currentEvidNumber))
{
const CEvidence *pEvidences[] = { pCurrentEvid };
parameter->UpdateStatisticsML(pEvidences, 1);
}
else
{
pCurrentInfEng->MarginalNodes(domain, DomainSize, 1);
const CPotential * pMargPot = pCurrentInfEng->GetQueryJPD();
parameter ->UpdateStatisticsEM(pMargPot, pCurrentEvid);
}
}
else
{
const CEvidence *pEvidences[] = { pCurrentEvid };
parameter->UpdateStatisticsML(pEvidences, 1);
}
}
itsML = itsML || !infIsNeed;
}
for(domainNodes = 0; domainNodes < numberOfParameters; domainNodes++ )
{
parameter = pGrModel->GetFactor(domainNodes);
CNumericDenseMatrix<float> *matForSending;
int matDim;
const int *pMatRanges;
int dataLength;
const float *pDataForSending;
matForSending = static_cast<CNumericDenseMatrix<float>*>
((parameter->GetDistribFun())->GetStatisticalMatrix(stMatTable));
matForSending->GetRanges(&matDim, &pMatRanges);
matForSending->GetRawData(&dataLength, &pDataForSending);
float *pDataRecv = new float[dataLength];
float *pDataRecv_copy = new float[dataLength];
MPI_Status status;
MPI_Allreduce((void*)pDataForSending, pDataRecv, dataLength, MPI_FLOAT, MPI_SUM,
MPI_COMM_WORLD);
CNumericDenseMatrix<float> *RecvMatrix =
static_cast<CNumericDenseMatrix<float>*>
(parameter->GetDistribFun()->GetStatisticalMatrix(stMatTable));
int dataLength_new;
float *pData_new;
RecvMatrix->GetRawData(&dataLength_new, (const float**)(&pData_new));
for(int t=0;t<dataLength_new;t++)
pData_new[t]=pDataRecv[t];
}
switch (pGrModel->GetModelType())
{
case mtBNet:
{
loglikOld = loglik;
loglik = 0.0f;
for(domainNodes = 0; domainNodes < numberOfParameters; domainNodes++)
{
parameter = pGrModel->GetFactor(domainNodes);
loglik += parameter->ProcessingStatisticalData(m_numberOfAllEvidences);
}
break;
}
case mtMRF2:
case mtMNet:
{
loglikOld = loglik;
loglik = _LearnPotentials();
break;
}
default:
{
PNL_THROW(CBadConst, "model type")
break;
}
}
stopExpression =
float(fabs(2 * (loglikOld - loglik) / (loglikOld + loglik)));
exit = ((stopExpression > epsilon) && (iteration <= GetMaxIterEM())) && !itsML;
if(exit)
{
ClearStatisticData();
}
delete pCurrentInfEng;
pCurrentInfEng = NULL;
}while(exit);
if(iteration > GetMaxIterEM())
{
PNL_THROW(CNotConverged, "maximum number of iterations")
}
SetNumProcEv( GetNumEv() );
}
int testShrinkObservedNodes()
{
int i/*,j*/;
int ret = TRS_OK;
/*prepare to read the values from console*/
EDistributionType dt;
int disType = -1;
EFactorType pt;
int paramType = -1;
/*read int disType corresponding DistributionType*/
while((disType<0)||(disType>0))/*now we have only Tabulars&Gaussian*/
{
trsiRead( &disType, "0", "DistributionType");
}
/*read int paramType corresponding FactorType*/
while((paramType<0)||(paramType>2))
{
trsiRead( ¶mType, "0", "FactorType");
}
dt = EDistributionType(disType);
pt = EFactorType(paramType);
int numberOfNodes = 0;
/*read number of nodes in Factor domain*/
while(numberOfNodes<=0)
{
trsiRead( &numberOfNodes, "1", "Number of Nodes in domain");
}
int numNodeTypes = 0;
/*read number of node types in model*/
while(numNodeTypes<=0)
{
trsiRead( &numNodeTypes, "1", "Number of node types in Domain");
}
//int seed1 = pnlTestRandSeed()/*%100000*/;
/*create string to display the value*/
/* char *value = new char[20];
value = _itoa(seed1, value, 10);
trsiRead(&seed1, value, "Seed for srand to define NodeTypes etc.");
delete []value;
trsWrite(TW_CON|TW_RUN|TW_DEBUG|TW_LST, "seed for rand = %d\n", seed1);
int *domain = (int *)trsGuardcAlloc(numberOfNodes, sizeof(int));
CNodeType * allNodeTypes = (CNodeType*)trsGuardcAlloc(numNodeTypes,
sizeof(CNodeType));
//To generate the NodeTypes we use rand()% and creates only Tabular now
for(i=0; i<numNodeTypes; i++)
{
allNodeTypes[i] = CNodeType(1, 1+rand()%(numNodeTypes+3));
}
*/
/*load data for parameter::ShrinkObservedNodes from console*/
intVector domain;
domain.assign( numberOfNodes, 0 );
nodeTypeVector allNodeTypes;
allNodeTypes.assign( numNodeTypes, CNodeType() );
/*read node types*/
for(i=0; i < numNodeTypes; i++)
{
int IsDiscrete = -1;
int NodeSize = -1;
while((IsDiscrete<0)||(IsDiscrete>1))
/*now we have tabular & Gaussian nodes!! */
trsiRead(&IsDiscrete, "1", "Is the node discrete?");
while(NodeSize<0)
trsiRead(&NodeSize, "2", "NodeSize of node");
allNodeTypes[i] = CNodeType( IsDiscrete != 0, NodeSize );
}
const CNodeType **nodeTypesOfDomain = (const CNodeType**)
trsGuardcAlloc(numberOfNodes, sizeof(CNodeType*));
int numData = 1;
int *Ranges = (int*)trsGuardcAlloc(numberOfNodes, sizeof(int));
/*associate nodes to node types*/
for(i=0; i<numberOfNodes; i++)
{
domain[i] = i;
int nodeAssociationToNodeType = -1;
while((nodeAssociationToNodeType<0)||(nodeAssociationToNodeType>=
numNodeTypes))
trsiRead(&nodeAssociationToNodeType, "0",
"node i has type nodeAssociationToNodeType");
nodeTypesOfDomain[i] = &allNodeTypes[nodeAssociationToNodeType];
// nodeTypesOfDomain[i] = &allNodeTypes[rand()%numNodeTypes];
Ranges[i] = nodeTypesOfDomain[i]->GetNodeSize();
numData=numData*Ranges[i];
}
CModelDomain* pMD = CModelDomain::Create( allNodeTypes, domain );
/*create factor according all information*/
CFactor *pMyParam = NULL;
float *data = (float *)trsGuardcAlloc(numData, sizeof(float));
char *stringVal;/* = (char*)trsGuardcAlloc(50, sizeof(char));*/
double val=0;
/*read the values from console*/
if(pt == ftPotential)
{
pMyParam = CTabularPotential::Create( &domain.front(), numberOfNodes, pMD );
/*here we can create data by multiply on 0.1 - numbers are nonnormalized*/
for(i=0; i<numData; i++)
{
val = 0.1*i;
stringVal = trsDouble(val);
trsdRead(&val, stringVal, "value of i's data position");
data[i] = (float)val;
//data[i] = (float)rand()/1000;
}
}
else
{
/*we can only read data from console - it must be normalized!!
(according their dimensions) - or we can normalize it by function!*/
if(pt == ftCPD)
pMyParam = CTabularCPD::Create( &domain.front(), numberOfNodes, pMD );
for(i=0; i<numData; i++)
{
val = -1;
while((val<0)||(val>1))
{
trsdRead(&val, "-1", "value of (2*i)'s data position");
}
data[i] = (float)val;
}
}
//trsWrite(TW_CON|TW_RUN|TW_DEBUG|TW_LST, "data for Factor = %d\n", data[i]);
pMyParam->AllocMatrix(data,matTable);
int nObsNodes = 0; /*rand()%numberOfNodes;*/
while((nObsNodes<=0)||(nObsNodes>numberOfNodes))
{
trsiRead(&nObsNodes, "1", "Number of Observed Nodes");
}
intVector myHelpForEvidence = intVector(domain.begin(), domain.end() );
int *ObsNodes = (int *)trsGuardcAlloc(nObsNodes, sizeof(int));
valueVector TabularValues;
TabularValues.assign( nObsNodes, (Value)0 );
char *strVal;
for(i=0; i<nObsNodes; i++)
{
//fixme - we need to have noncopy only different ObsNodes
/* j = rand()%(numberOfNodes-i);*/
int numberOfObsNode = -1;
strVal = trsInt(i);
intVector::iterator j = std::find( myHelpForEvidence.begin(), myHelpForEvidence.end(), numberOfObsNode );
while((numberOfObsNode<0)||(numberOfObsNode>numberOfNodes)||
(j==myHelpForEvidence.end()))
{
trsiRead(&numberOfObsNode, strVal,"Number of i's observed node");
j = std::find(myHelpForEvidence.begin(), myHelpForEvidence.end(),
numberOfObsNode);
}
//ObsNodes[i] = myHelpForEvidence[j];
myHelpForEvidence.erase( j );
ObsNodes[i] = numberOfObsNode;
int valueOfNode = -1;
int maxValue = (*nodeTypesOfDomain[ObsNodes[i]]).GetNodeSize();
while((valueOfNode<0)||(valueOfNode>=maxValue))
{
trsiRead(&valueOfNode,"0","this is i's observed node value");
}
TabularValues[i].SetInt(valueOfNode);
/*rand()%((*nodeTypesOfDomain[ObsNodes[i]]).pgmGetNodeSize());*/
}
CEvidence* pEvidence = CEvidence::Create( pMD, nObsNodes, ObsNodes, TabularValues );
myHelpForEvidence.clear();
CNodeType *ObservedNodeType = (CNodeType*)trsGuardcAlloc(1,
sizeof(CNodeType));
*ObservedNodeType = CNodeType(1,1);
CPotential *myTakedInFactor = static_cast<CPotential*>(pMyParam)->ShrinkObservedNodes(pEvidence);
const int *myfactorDomain;
int factorDomSize ;
myTakedInFactor->GetDomain(&factorDomSize, &myfactorDomain);
#if 0
CNumericDenseMatrix<float> *mySmallMatrix = static_cast<
CNumericDenseMatrix<float>*>(myTakedInFactor->GetMatrix(matTable));
int n;
const float* mySmallData;
mySmallMatrix->GetRawData(&n, &mySmallData);
int nDims; // = mySmallMatrix->GetNumberDims();
const int * mySmallRanges;
mySmallMatrix->GetRanges(&nDims, &mySmallRanges);
if(nDims!=numberOfNodes)
{
ret = TRS_FAIL;
trsWrite(TW_CON|TW_RUN|TW_DEBUG|TW_LST, "nDims = %d\n", nDims);
}
else
{
int numSmallData = 1;
for(i=0; i<nDims; i++)
{
numSmallData = numSmallData*mySmallRanges[i];
trsWrite(TW_CON|TW_RUN|TW_DEBUG|TW_LST, "Range[%d] = %d\n", i,
mySmallRanges[i]);
}
for(i=0; i<numSmallData; i++)
{
trsWrite(TW_CON|TW_RUN|TW_DEBUG|TW_LST, "mySmallData[%d] = %f ",
i, mySmallData[i]);
}
}
#endif
//getchar();
delete(myTakedInFactor);
delete (pMyParam);
delete pMD;
//test gaussian parameter
nodeTypeVector nTypes;
nTypes.assign( 2, CNodeType() );
nTypes[0] = CNodeType( 0, 2 );
nTypes[1] = CNodeType( 0,1 );
intVector domn = intVector(3,0);
domn[1] = 1;
domn[2] = 1;
CModelDomain* pMD1 = CModelDomain::Create( nTypes, domn );
domn[2] = 2;
CPotential *BigFactor = CGaussianPotential::CreateUnitFunctionDistribution(
&domn.front(), domn.size(), pMD1,0 );
float mean[] = { 1.0f, 3.2f};
CPotential *SmallDelta = CGaussianPotential::CreateDeltaFunction( &domn.front(), 1, pMD1, mean, 1 );
domn.resize( 2 );
domn[0] = 1;
domn[1] = 2;
CPotential *SmallFunct = CGaussianPotential::Create( &domn.front(),
domn.size(), pMD1);
float datH[] = { 1.1f, 2.2f, 3.3f };
float datK[] = { 1.2f, 2.3f, 2.3f, 3.4f, 5.6f, 6.7f, 3.4f, 6.7f, 9.0f };
SmallFunct->AllocMatrix( datH, matH );
SmallFunct->AllocMatrix( datK, matK );
static_cast<CGaussianPotential*>(SmallFunct)->SetCoefficient( 0.2f, 1 );
CPotential* multFact = BigFactor->Multiply( SmallDelta );
CPotential* nextMultFact = multFact->Multiply( SmallFunct );
domn[0] = 0;
domn[1] = 1;
CPotential *marginalized = static_cast<CPotential*>(nextMultFact->Marginalize( &domn.front(), domn.size() ));
int isSpecific = marginalized->IsDistributionSpecific();
if( isSpecific )
{
trsWrite(TW_CON|TW_RUN|TW_DEBUG|TW_LST, "\nGaussian Distribution is specific");
}
delete BigFactor;
delete SmallFunct;
delete SmallDelta;
delete pMD1;
int ranges_memory_flag = trsGuardCheck(Ranges);
int data_memory_flag = trsGuardCheck(data);
int nodeTypesOfDomain_mem_b = trsGuardCheck(nodeTypesOfDomain);
int ObsNodes_mem_b = trsGuardCheck(ObsNodes);
int ObsNodeType_mem_b = trsGuardCheck(ObservedNodeType);
if(((ranges_memory_flag)||(data_memory_flag)||
(nodeTypesOfDomain_mem_b)||
(ObsNodes_mem_b)||(ObsNodeType_mem_b)))
{
ret = TRS_FAIL;
return trsResult( ret, ret == TRS_OK ? "No errors" :
"Bad test on ShrinkObservedNodes Method - memory");
}
else
{
trsGuardFree(ObservedNodeType);
trsGuardFree(ObsNodes);
trsGuardFree(nodeTypesOfDomain);
trsGuardFree(data);
trsGuardFree(Ranges);
}
return trsResult( ret, ret == TRS_OK ? "No errors" :
"Bad test on ShrinkObservedNodes Method");
}
void CParEMLearningEngine::LearnContMPI()
{
CStaticGraphicalModel *pGrModel = this->GetStaticModel();
PNL_CHECK_IS_NULL_POINTER(pGrModel);
PNL_CHECK_LEFT_BORDER(GetNumEv() - GetNumberProcEv() , 1);
CInfEngine *pInfEng = NULL;
pInfEng = CJtreeInfEngine::Create(pGrModel);
float loglik = 0.0f;
int domainNodes;
CFactor *parameter = NULL;
int numberOfParameters = pGrModel->GetNumberOfParameters();
int nFactors = pGrModel->GetNumberOfFactors();
const CEvidence *pEv;
CFactor *pFactor;
int iteration = 0;
int ev;
int i,numSelfEvidences,NumberOfProcesses, MyRank;
int start_mpi, finish_mpi;
MPI_Comm_size(MPI_COMM_WORLD, &NumberOfProcesses);
MPI_Comm_rank(MPI_COMM_WORLD, &MyRank);
if (IsAllObserved())
{
int i;
float **evid = NULL;
EDistributionType dt;
CFactor *factor = NULL;
for (i = 0; i < nFactors; i++)
{
factor = pGrModel->GetFactor(i);
factor->UpdateStatisticsML(&m_Vector_pEvidences[GetNumberProcEv()],
GetNumEv() - GetNumberProcEv());
}
m_critValue.push_back(UpdateModel());
}
else
{
bool bContinue;
const CPotential * pot;
do
{
ClearStatisticData();
iteration++;
numSelfEvidences = (GetNumEv() - GetNumberProcEv()) / NumberOfProcesses;
start_mpi = GetNumberProcEv() + numSelfEvidences * MyRank;
if (MyRank < NumberOfProcesses - 1)
finish_mpi = start_mpi + numSelfEvidences;
else
finish_mpi = GetNumEv();
for(int ev = start_mpi; ev < finish_mpi; ev++)
{
bool bInfIsNeed = !GetObsFlags(ev)->empty();
pEv = m_Vector_pEvidences[ev];
if( bInfIsNeed )
{
pInfEng->EnterEvidence(pEv, 0, 0);
}
int i;
for( i = 0; i < nFactors; i++ )
{
pFactor = pGrModel->GetFactor(i);
int nnodes;
const int * domain;
pFactor->GetDomain( &nnodes, &domain );
if( bInfIsNeed && !IsDomainObserved(nnodes, domain, ev ) )
{
pInfEng->MarginalNodes( domain, nnodes, 1 );
pot = pInfEng->GetQueryJPD();
pFactor->UpdateStatisticsEM( /*pInfEng->GetQueryJPD */ pot, pEv );
}
else
{
pFactor->UpdateStatisticsML( &pEv, 1 );
}
}
}
for(domainNodes = 0; domainNodes < numberOfParameters; domainNodes++ )
{
parameter = pGrModel->GetFactor(domainNodes);
C2DNumericDenseMatrix<float> *matMeanForSending;
C2DNumericDenseMatrix<float> *matCovForSending;
int dataLengthM,dataLengthC;
const float *pMeanDataForSending;
const float *pCovDataForSending;
matMeanForSending = static_cast<C2DNumericDenseMatrix<float>*>
((parameter->GetDistribFun())->GetStatisticalMatrix(stMatMu));
matMeanForSending->GetRawData(&dataLengthM, &pMeanDataForSending);
matCovForSending = static_cast<C2DNumericDenseMatrix<float>*>
((parameter->GetDistribFun())->GetStatisticalMatrix(stMatSigma));
matCovForSending->GetRawData(&dataLengthC, &pCovDataForSending);
float *pMeanDataRecv = new float[dataLengthM];
float *pCovDataRecv = new float[dataLengthC];
MPI_Status status;
MPI_Allreduce((void*)pMeanDataForSending, pMeanDataRecv, dataLengthM, MPI_FLOAT, MPI_SUM,
MPI_COMM_WORLD);
MPI_Allreduce((void*)pCovDataForSending, pCovDataRecv, dataLengthC, MPI_FLOAT, MPI_SUM,
MPI_COMM_WORLD);
memcpy((void*)pMeanDataForSending,pMeanDataRecv,dataLengthM*sizeof(float));
memcpy((void*)pCovDataForSending,pCovDataRecv,dataLengthC*sizeof(float));
}
loglik = UpdateModel();
if( GetMaxIterEM() != 1)
{
bool flag = iteration == 1 ? true :
(fabs(2*(m_critValue.back()-loglik)/(m_critValue.back() + loglik)) > GetPrecisionEM() );
bContinue = GetMaxIterEM() > iteration && flag;
}
else
{
bContinue = false;
}
m_critValue.push_back(loglik);
}while(bContinue);
}
SetNumProcEv( GetNumEv() );
}
CBNet* CreateBNet()
{
// Creation Water-Sprinkler Bayesian network
const int numOfNds = 4;//*<-
// 1 STEP:
// need to specify the graph structure of the model;
// there are two way to do it
CGraph *pGraph;
// Graph creation using neighbors list
int numOfNbrs[numOfNds] = { 2, 2, 2, 2 };//*<-
int nbrs0[] = { 1, 2 };//*<-
int nbrs1[] = { 0, 3 };//*<-
int nbrs2[] = { 0, 3 };//*<-
int nbrs3[] = { 1, 2 };//*<-
// number of neighbors for every node
int *nbrs[] = { nbrs0, nbrs1, nbrs2, nbrs3 };//*<-
// neighbors can be of either one of the three following types:
// a parent, a child (for directed arcs) or just a neighbor (for undirected graphs).
// Accordingly, the types are ntParent, ntChild or ntNeighbor.
ENeighborType nbrsTypes0[] = { ntChild, ntChild };//*<-
ENeighborType nbrsTypes1[] = { ntParent, ntChild };//*<-
ENeighborType nbrsTypes2[] = { ntParent, ntChild };//*<-
ENeighborType nbrsTypes3[] = { ntParent, ntParent };//*<-
ENeighborType *nbrsTypes[] = { nbrsTypes0, nbrsTypes1,nbrsTypes2, nbrsTypes3 };//*<-
// this is creation of a directed graph for the BNet model using neighbors list
pGraph = CGraph::Create( numOfNds, numOfNbrs, nbrs, nbrsTypes );
// 2 STEP:
// Creation NodeType objects and specify node types for all nodes of the model.
nodeTypeVector nodeTypes;
// number of node types is 1, because all nodes are of the same type
// all four are discrete and binary
CNodeType nt(1,2);//*<-
nodeTypes.push_back(nt);
intVector nodeAssociation;
// reflects association between node numbers and node types
// nodeAssociation[k] is a number of node type object in the
// node types array for the k-th node
nodeAssociation.assign(numOfNds, 0);
// 2 STEP:
// Creation base for BNet using Graph, types of nodes and nodes association
CBNet* pBNet = CBNet::Create( numOfNds, nodeTypes, nodeAssociation, pGraph );
// 3 STEP:
// Allocation space for all factors of the model
pBNet->AllocFactors();
// 4 STEP:
// Creation factors and attach their to model
//create raw data tables for CPDs
float table0[] = { 0.5f, 0.5f };//*<-
float table1[] = { 0.5f, 0.5f, 0.9f, 0.1f };
float table2[] = { 0.8f, 0.2f, 0.2f, 0.8f };
float table3[] = { 1.0f, 0.0f, 0.1f, 0.9f, 0.1f, 0.9f, 0.01f, 0.99f };
float* table[] = { table0, table1, table2, table3 };//*<-
int i;
for( i = 0; i < numOfNds; ++i )
{
pBNet->AllocFactor(i);
CFactor* pFactor = pBNet->GetFactor(i);
pFactor->AllocMatrix( table[i], matTable );
}
return pBNet;
}
void CStaticStructLearnSEM::CreateNeighborCPDs(CBNet* pBNet,
pCPDVector* vNeighborCPDs, EDGEOPVECTOR* vValidMoves, intVector* RevCorrespDel)
{
CGraph* pGraph = pBNet->GetGraph();
CDAG* pDAG = CDAG::Create(*pGraph);
CModelDomain* pMD = pBNet->GetModelDomain();
intVector vDiscrete, vContinuous;
intVector vAncestor, vDescent;
intVector vMixture, vMix;
const CNodeType* nt;
CFactor* factor;
int i, j, position;
vAncestor.assign(m_vAncestor.begin(), m_vAncestor.end());
vDescent.assign(m_vDescent.begin(), m_vDescent.end());
pBNet->FindMixtureNodes(&vMix);
for(i=0; i<vMix.size(); i++)
{
factor = pBNet->GetFactor(vMix[i]);
j = static_cast<CMixtureGaussianCPD*>(factor) -> GetNumberOfMixtureNode();
vMixture.push_back(j);
}
for(i=0; i<m_nNodes; i++)
{
nt = pMD->GetVariableType(i);
if( nt->IsDiscrete() )
{
vDiscrete.push_back(i);
}
else
vContinuous.push_back(i);
}
vValidMoves->clear();
vNeighborCPDs->clear();
RevCorrespDel->clear();
pDAG->GetAllValidMove(vValidMoves, &vMixture.front(), vMixture.size(), m_nMaxFanIn,
&vDiscrete, &vContinuous, &vDescent, &vAncestor );
int nMoves = vValidMoves->size();
intVector domain;
EDGEOP curMove;
int start, end;
for(i=0; i<nMoves; i++)
{
domain.clear();
curMove = (*vValidMoves)[i];
switch (curMove.DAGChangeType)
{
case DAG_DEL :
start = curMove.originalEdge.startNode;
end = curMove.originalEdge.endNode;
factor = pBNet->GetFactor(end);
factor->GetDomain(&domain);
position = std::find(domain.begin(), domain.end(), start)
- domain.begin();
domain.erase(domain.begin()+position);
vNeighborCPDs->push_back(CreateRandomCPD(domain.size(), &domain.front(), pBNet));
break;
case DAG_ADD :
start = curMove.originalEdge.startNode;
end = curMove.originalEdge.endNode;
factor = pBNet->GetFactor(end);
factor->GetDomain(&domain);
domain.insert(domain.begin(), start);
vNeighborCPDs->push_back(CreateRandomCPD(domain.size(), &domain.front(), pBNet));
break;
case DAG_REV :
end = curMove.originalEdge.startNode;
start = curMove.originalEdge.endNode;
factor = pBNet->GetFactor(end);
factor->GetDomain(&domain);
domain.insert(domain.begin(), start);
vNeighborCPDs->push_back(CreateRandomCPD(domain.size(), &domain.front(), pBNet));
break;
}
}
RevCorrespDel->assign(nMoves, -1);
EDGEOP pre_move;
for(i=0; i<nMoves; i++)
{
curMove = (*vValidMoves)[i];
if(curMove.DAGChangeType == DAG_REV)
{
start = curMove.originalEdge.startNode;
end = curMove.originalEdge.endNode;
for(j=0; j<nMoves; j++)
{
pre_move = (*vValidMoves)[j];
if( (start == pre_move.originalEdge.startNode) &&
(end == pre_move.originalEdge.endNode) &&
(pre_move.DAGChangeType == DAG_DEL) )
{
(*RevCorrespDel)[i] = j;
break;
}
}
}
}
}
CFactor* CFactor::CopyWithNewDomain(const CFactor *factor, intVector &domain,
CModelDomain *pMDNew,
const intVector& /* obsIndices */)
{
int domSize = domain.size();
intVector domOld;
factor->GetDomain( &domOld );
if( int(domOld.size()) != domSize )
{
PNL_THROW( CBadArg, "number of nodes" );
}
CModelDomain *pMDOld = factor->GetModelDomain();
//check is the types are the same
const pConstNodeTypeVector* ntFactor = factor->GetArgType();
/*
const CNodeType *nt;
for( int i = 0; i < domSize; i++ )
{
nt = (*ntFactor)[i];
if( nt->IsDiscrete() )
{
if( nt->GetNodeSize() == 1 )
{
if( *pMDOld->GetVariableType(domOld[i]) !=
*pMDNew->GetVariableType(domain[i]))
{
PNL_THROW(CInconsistentType, "types of variables should correspond");
}
}
else
{
if( *nt != *pMDNew->GetVariableType(domain[i]) )
{
PNL_THROW(CInconsistentType, "types of variables should correspond");
}
}
}
else
{
if( nt->GetNodeSize() == 0 )
{
if( *pMDOld->GetVariableType(domOld[i]) !=
*pMDNew->GetVariableType(domain[i]))
{
PNL_THROW(CInconsistentType, "types of variables should correspond");
}
}
else
{
if( *nt != *pMDNew->GetVariableType(domain[i]) )
{
PNL_THROW(CInconsistentType, "types of variables should correspond");
}
}
}
}
*/
const CNodeType *nt;
int i;
intVector obsPositions;
factor->GetObsPositions(&obsPositions);
if( obsPositions.size() )
{
intVector::iterator iterEnd = obsPositions.end();
for( i = 0; i < domSize; i++)
{
if( std::find( obsPositions.begin(),iterEnd, i) != iterEnd )
{
if( *pMDOld->GetVariableType(domOld[i]) !=
*pMDNew->GetVariableType(domain[i]))
{
PNL_THROW(CInconsistentType, "types of variables should correspond");
}
}
}
}
else
{
for( i = 0; i < domSize; i++ )
{
nt = (*ntFactor)[i];
if( *nt != *pMDNew->GetVariableType(domain[i]) )
{
PNL_THROW(CInconsistentType, "types of variables should correspond");
}
}
}
CFactor *pNewFactor;
switch ( factor->GetFactorType() )
{
case ftPotential:
{
switch ( factor->GetDistributionType() )
{
case dtTabular:
{
pNewFactor = CTabularPotential::
Copy(static_cast<const CTabularPotential*>(factor) );
break;
}
case dtGaussian:
{
pNewFactor = CGaussianPotential::
Copy(static_cast<const CGaussianPotential*>(factor));
break;
}
case dtScalar:
{
pNewFactor = CScalarPotential::
Copy( static_cast<const CScalarPotential*>(factor) );
break;
}
default:
{
PNL_THROW(CNotImplemented, "distribution type" );
}
}
break;
}
case ftCPD:
{
switch ( factor->GetDistributionType() )
{
case dtTabular:
{
pNewFactor = CTabularCPD::
Copy(static_cast<const CTabularCPD*>(factor));
break;
}
case dtGaussian:
{
pNewFactor = CGaussianCPD::
Copy(static_cast<const CGaussianCPD*>(factor));
break;
}
case dtCondGaussian:
{
pNewFactor = CGaussianCPD::
Copy(static_cast<const CGaussianCPD*>(factor));
break;
}
case dtSoftMax:
{
pNewFactor = CSoftMaxCPD::
Copy(static_cast<const CSoftMaxCPD*>(factor));
break;
}
case dtCondSoftMax:
{
pNewFactor = CSoftMaxCPD::
Copy(static_cast<const CSoftMaxCPD*>(factor));
break;
}
case dtMixGaussian:
{
pNewFactor = CMixtureGaussianCPD::Copy(
static_cast<const CMixtureGaussianCPD*>(factor));
break;
}
default:
{
PNL_THROW(CNotImplemented, "distribution type" );
}
}
break;
}
default:
{
PNL_THROW(CNotImplemented, "factor type" );
}
}
PNL_CHECK_IF_MEMORY_ALLOCATED(pNewFactor);
/*
if( pMDNew == factor->GetModelDomain())
{
return pNewFactor;
}
else*/
{
pNewFactor->m_Domain = intVector(domain);
pNewFactor->SetModelDomain(pMDNew, 0);
return pNewFactor;
}
}
CExInfEngine< INF_ENGINE, MODEL, FLAV, FALLBACK_ENGINE1, FALLBACK_ENGINE2 >::CExInfEngine( CStaticGraphicalModel const *gm )
: CInfEngine( itEx, gm ), evidence_mine( false ),
maximize( 0 ), MPE_ev( 0 ), query_JPD( 0 ), graphical_model( gm )
{
int i, j, k;
intVector dom;
intVector conv;
CFactor *fac;
PNL_MAKE_LOCAL( CGraph *, gr, gm, GetGraph() );
PNL_MAKE_LOCAL( int, sz, gr, GetNumberOfNodes() );
gr->GetConnectivityComponents( &decomposition );
for ( i = decomposition.size(); i--; )
{
std::sort( decomposition[i].begin(), decomposition[i].end() );
}
if ( PNL_IS_EXINFENGINEFLAVOUR_UNSORTED( FLAV ) )
{
gr->GetTopologicalOrder( &conv );
}
orig2comp.resize( sz );
orig2idx.resize( sz );
for ( k = 2; k--; )
{
for ( i = decomposition.size(); i--; )
{
for ( j = decomposition[i].size(); j--; )
{
orig2comp[decomposition[i][j]] = i;
orig2idx[decomposition[i][j]] = j;
}
}
if ( PNL_IS_EXINFENGINEFLAVOUR_UNSORTED( FLAV ) && k )
{
for ( i = sz; i--; )
{
decomposition[orig2comp[conv[i]]][orig2idx[conv[i]]] = i;
}
}
else
{
break;
}
}
graphs.resize( decomposition.size() );
models.resize( decomposition.size() );
engines.resize( decomposition.size() );
for ( i = decomposition.size(); i--; )
{
graphs[i] = gr->ExtractSubgraph( decomposition[i] );
#if 0
std::cout << "graph " << i << std::endl;
graphs[i]->Dump();
#endif
}
node_types.resize( decomposition.size() );
node_assoc.resize( decomposition.size() );
for ( i = 0, k = 0; i < decomposition.size(); ++i )
{
node_types[i].resize( decomposition[i].size() );
node_assoc[i].resize( decomposition[i].size() );
for ( j = 0; j < decomposition[i].size(); ++j )
{
node_types[i][j] = *gm->GetNodeType( decomposition[i][j] );
node_assoc[i][j] = j;
}
}
for ( i = decomposition.size(); i--; )
{
models[i] = MODEL::Create( decomposition[i].size(), node_types[i], node_assoc[i], graphs[i] );
}
for ( i = 0; i < gm->GetNumberOfFactors(); ++i )
{
fac = gm->GetFactor( i );
fac->GetDomain( &dom );
#if 0
std::cout << "Ex received orig factor" << std::endl;
fac->GetDistribFun()->Dump();
#endif
k = orig2comp[dom[0]];
for ( j = dom.size(); j--; )
{
dom[j] = orig2idx[dom[j]];
}
fac = CFactor::CopyWithNewDomain( fac, dom, models[k]->GetModelDomain() );
#if 0
std::cout << "Ex mangled it to" << std::endl;
fac->GetDistribFun()->Dump();
#endif
models[k]->AttachFactor( fac );
}
for ( i = decomposition.size(); i--; )
{
switch ( decomposition[i].size() )
{
case 1:
engines[i] = FALLBACK_ENGINE1::Create( models[i] );
continue;
case 2:
engines[i] = FALLBACK_ENGINE2::Create( models[i] );
continue;
default:
engines[i] = INF_ENGINE::Create( models[i] );
}
}
}
void CEMLearningEngine::LearnExtraCPDs(int nMaxFamily, pCPDVector* additionalCPDs, floatVector* additionalLLs)
{
CStaticGraphicalModel *pGrModel = this->GetStaticModel();
PNL_CHECK_IS_NULL_POINTER(pGrModel);
PNL_CHECK_LEFT_BORDER(GetNumEv(), 1);
int numberOfFactors = pGrModel->GetNumberOfFactors();
int numberOfAddFactors = additionalCPDs->size();
additionalLLs->resize(numberOfAddFactors);
additionalLLs->clear();
m_vFamilyLogLik.resize(numberOfFactors);
float loglik = 0.0f, ll;
int i, ev;
int iteration = 0;
const CEvidence* pEv;
CFactor *factor = NULL;
int nnodes;
const int * domain;
bool bInfIsNeed;
CInfEngine *pInfEng = m_pInfEngine;
if (IsAllObserved())
{
for (i = 0; i < numberOfFactors; i++)
{
factor = pGrModel->GetFactor(i);
factor->UpdateStatisticsML(&m_Vector_pEvidences[GetNumberProcEv()],
GetNumEv() - GetNumberProcEv());
}
for( ev = 0; ev < GetNumEv() ; ev++)
{
pEv = m_Vector_pEvidences[ev];
for( i = 0; i < numberOfAddFactors; i++ )
{
factor = static_cast<CFactor*>((*additionalCPDs)[i]);
factor->UpdateStatisticsML( &pEv, 1 );
}
}
switch (pGrModel->GetModelType())
{
case mtBNet:
{
for( i = 0; i<numberOfFactors; i++ )
{
factor = pGrModel->GetFactor(i);
ll = factor->ProcessingStatisticalData( GetNumEv());
m_vFamilyLogLik[i] = ll;
loglik += ll;
}
for( i = 0; i < numberOfAddFactors; i++ )
{
factor = static_cast<CFactor*>((*additionalCPDs)[i]);
ll = factor->ProcessingStatisticalData( GetNumEv());
(*additionalLLs)[i] = ll;
}
break;
}
case mtMRF2:
case mtMNet:
{
break;
}
default:
{
PNL_THROW(CBadConst, "model type" )
break;
}
}
m_critValue.push_back(loglik);
}
else
{
int CBICLearningEngine::DimOfModel(const CStaticGraphicalModel *pModel)
{
/*
compute dimension of the model in (d)
it using in BIC criterion:
BIC = LogLic - 0.5*d*log(N)
*/
int nParam = pModel->GetNumberOfFactors();
CFactor *param = NULL;
int dimOfModel = 0;
int dim = 1;
CMatrix<float> *matrix;;
for (int i = 0; i < nParam; i++)
{
dim = 1;
param = pModel->GetFactor(i);
switch (param->GetFactorType())
{
case ftCPD:
{
switch (param->GetDistributionType())
{
case dtTabular:
{
matrix = param->GetMatrix(matTable);
int size;
const int *ranges;
static_cast<CNumericDenseMatrix<float>*>(matrix)->
GetRanges(&size, &ranges);
for(int j=0; j < size - 1; j++)
{
dim *= ranges[j];
}
dim *= ranges[size-1]-1;
break;
}//case dtTabular
case dtGaussian:
{
PNL_THROW(CNotImplemented,"Gaussian")
break;
}//case dtGaussian
case dtCondGaussian:
{
PNL_THROW(CNotImplemented,"CondGaussian")
break;
}//case dtCondGaussian
default:
{
PNL_THROW(CBadConst,"distribution type")
break;
}
}//swith(param->GetFactorType)
break;
}//end case ftCPD
case ftPotential:
{
PNL_THROW(CNotImplemented,"Factor")
break;
}
default:
{
PNL_THROW(CBadConst,"FactorType")
break;
}
}//end switch(param->GetFactor)
dimOfModel += dim;
}//end for(i)
return dimOfModel;
}
void CEMLearningEngine::Learn()
{
CStaticGraphicalModel *pGrModel = this->GetStaticModel();
PNL_CHECK_IS_NULL_POINTER(pGrModel);
PNL_CHECK_LEFT_BORDER(GetNumEv() - GetNumberProcEv() , 1);
CInfEngine *pInfEng = NULL;
if (m_pInfEngine)
{
pInfEng = m_pInfEngine;
}
else
{
if (!m_bAllObserved)
{
pInfEng = CJtreeInfEngine::Create(pGrModel);
m_pInfEngine = pInfEng;
}
}
float loglik = 0.0f;
int nFactors = pGrModel->GetNumberOfFactors();
const CEvidence *pEv;
CFactor *pFactor;
int iteration = 0;
int ev;
bool IsCastNeed = false;
int i;
for( i = 0; i < nFactors; i++ )
{
pFactor = pGrModel->GetFactor(i);
EDistributionType dt = pFactor->GetDistributionType();
if ( dt == dtSoftMax ) IsCastNeed = true;
}
float ** full_evid = NULL;
if (IsCastNeed)
{
BuildFullEvidenceMatrix(&full_evid);
}
if (IsAllObserved())
{
int i;
float **evid = NULL;
EDistributionType dt;
CFactor *factor = NULL;
for (i = 0; i < nFactors; i++)
{
factor = pGrModel->GetFactor(i);
dt = factor->GetDistributionType();
if (dt != dtSoftMax)
{
factor->UpdateStatisticsML(&m_Vector_pEvidences[GetNumberProcEv()],
GetNumEv() - GetNumberProcEv());
}
else
{
intVector family;
family.resize(0);
pGrModel->GetGraph()->GetParents(i, &family);
family.push_back(i);
CSoftMaxCPD* SoftMaxFactor = static_cast<CSoftMaxCPD*>(factor);
SoftMaxFactor->BuildCurrentEvidenceMatrix(&full_evid,
&evid,family,m_Vector_pEvidences.size());
SoftMaxFactor->InitLearnData();
SoftMaxFactor->SetMaximizingMethod(m_MaximizingMethod);
SoftMaxFactor->MaximumLikelihood(evid, m_Vector_pEvidences.size(),
0.00001f, 0.01f);
SoftMaxFactor->CopyLearnDataToDistrib();
for (int k = 0; k < factor->GetDomainSize(); k++)
{
delete [] evid[k];
}
delete [] evid;
}
}
m_critValue.push_back(UpdateModel());
}
else
{
bool bContinue;
const CPotential * pot;
/* bool IsCastNeed = false;
int i;
for( i = 0; i < nFactors; i++ )
{
pFactor = pGrModel->GetFactor(i);
EDistributionType dt = pFactor->GetDistributionType();
if ( dt == dtSoftMax ) IsCastNeed = true;
}
float ** full_evid;
if (IsCastNeed)
{
BuildFullEvidenceMatrix(full_evid);
}*/
do
{
ClearStatisticData();
iteration++;
for( ev = GetNumberProcEv(); ev < GetNumEv() ; ev++ )
{
bool bInfIsNeed = !GetObsFlags(ev)->empty();
pEv = m_Vector_pEvidences[ev];
if( bInfIsNeed )
{
pInfEng->EnterEvidence(pEv, 0, 0);
}
int i;
for( i = 0; i < nFactors; i++ )
{
pFactor = pGrModel->GetFactor(i);
int nnodes;
const int * domain;
pFactor->GetDomain( &nnodes, &domain );
if( bInfIsNeed && !IsDomainObserved(nnodes, domain, ev ) )
{
pInfEng->MarginalNodes( domain, nnodes, 1 );
pot = pInfEng->GetQueryJPD();
if ( (!(m_Vector_pEvidences[ev])->IsNodeObserved(i)) && (IsCastNeed) )
{
Cast(pot, i, ev, &full_evid);
}
EDistributionType dt;
dt = pFactor->GetDistributionType();
if ( !(dt == dtSoftMax) )
pFactor->UpdateStatisticsEM( /*pInfEng->GetQueryJPD */ pot, pEv );
}
else
{
if ((pFactor->GetDistributionType()) != dtSoftMax)
pFactor->UpdateStatisticsML( &pEv, 1 );
}
}
}
int i;
/*
printf ("\n My Full Evidence Matrix");
for (i=0; i<nFactors; i++)
{
for (j=0; j<GetNumEv(); j++)
{
printf ("%f ", full_evid[i][j]);
}
printf("\n");
}
*/
float **evid = NULL;
EDistributionType dt;
CFactor *factor = NULL;
// int i;
for (i = 0; i < nFactors; i++)
{
factor = pGrModel->GetFactor(i);
dt = factor->GetDistributionType();
if (dt == dtSoftMax)
{
intVector family;
family.resize(0);
pGrModel->GetGraph()->GetParents(i, &family);
family.push_back(i);
CSoftMaxCPD* SoftMaxFactor = static_cast<CSoftMaxCPD*>(factor);
SoftMaxFactor->BuildCurrentEvidenceMatrix(&full_evid,
&evid,family,m_Vector_pEvidences.size());
SoftMaxFactor->InitLearnData();
SoftMaxFactor->SetMaximizingMethod(m_MaximizingMethod);
// SoftMaxFactor->MaximumLikelihood(evid, m_numberOfLastEvidences,
SoftMaxFactor->MaximumLikelihood(evid, m_Vector_pEvidences.size(),
0.00001f, 0.01f);
SoftMaxFactor->CopyLearnDataToDistrib();
for (int k = 0; k < factor->GetDomainSize(); k++)
{
delete [] evid[k];
}
delete [] evid;
}
}
loglik = UpdateModel();
if( GetMaxIterEM() != 1)
{
bool flag = iteration == 1 ? true :
(fabs(2*(m_critValue.back()-loglik)/(m_critValue.back() + loglik)) > GetPrecisionEM() );
bContinue = GetMaxIterEM() > iteration && flag;
}
else
{
bContinue = false;
}
m_critValue.push_back(loglik);
}while(bContinue);
}
SetNumProcEv( GetNumEv() );
if (IsCastNeed)
{
int NumOfNodes = pGrModel->GetGraph()->GetNumberOfNodes();
for (i=0; i<NumOfNodes; i++)
{
delete [] full_evid[i];
}
delete [] full_evid;
}
}
int testSetStatistics()
{
int ret = TRS_OK;
float eps = 0.1f;
int seed = pnlTestRandSeed();
pnlSeed( seed );
CBNet *pBNet = pnlExCreateCondGaussArBNet();
CModelDomain *pMD = pBNet->GetModelDomain();
CGraph *pGraph = CGraph::Copy(pBNet->GetGraph());
CBNet *pBNet1 = CBNet::CreateWithRandomMatrices( pGraph, pMD );
pEvidencesVector evidences;
int nEvidences = pnlRand( 3000, 4000);
pBNet->GenerateSamples( &evidences, nEvidences );
int i;
for( i = 0; i < nEvidences; i++)
{
//evidences[i]->MakeNodeHiddenBySerialNum(0);
}
CEMLearningEngine *pLearn = CEMLearningEngine::Create(pBNet1);
pLearn->SetData( nEvidences, &evidences.front() );
pLearn->SetMaxIterEM();
pLearn->Learn();
for( i = 0; i < pBNet->GetNumberOfFactors(); i++ )
{
if( ! pBNet->GetFactor(i)->IsFactorsDistribFunEqual(pBNet1->GetFactor(i), eps))
{
ret = TRS_FAIL;
pBNet->GetFactor(i)->GetDistribFun()->Dump();
pBNet1->GetFactor(i)->GetDistribFun()->Dump();
}
}
CDistribFun *pDistr;
const CMatrix<float>* pMat;
CFactor *pCPD;
pDistr = pBNet1->GetFactor(0)->GetDistribFun();
pMat = pDistr->GetStatisticalMatrix(stMatTable);
pCPD = pBNet->GetFactor(0);
pCPD->SetStatistics(pMat, stMatTable);
pCPD->ProcessingStatisticalData(nEvidences);
if( ! pCPD->IsFactorsDistribFunEqual(pBNet1->GetFactor(0), 0.0001f) )
{
ret = TRS_FAIL;
}
pDistr = pBNet1->GetFactor(1)->GetDistribFun();
int parentVal;
pCPD = pBNet->GetFactor(1);
parentVal = 0;
pCPD->SetStatistics(pMat, stMatCoeff);
pMat = pDistr->GetStatisticalMatrix(stMatMu, &parentVal);
pCPD->SetStatistics(pMat, stMatMu, &parentVal);
pMat = pDistr->GetStatisticalMatrix(stMatSigma, &parentVal);
pCPD->SetStatistics(pMat, stMatSigma, &parentVal);
parentVal = 1;
pMat = pDistr->GetStatisticalMatrix(stMatMu, &parentVal);
pCPD->SetStatistics(pMat, stMatMu, &parentVal);
pMat = pDistr->GetStatisticalMatrix(stMatSigma, &parentVal);
pCPD->SetStatistics(pMat, stMatSigma, &parentVal);
pCPD->ProcessingStatisticalData(nEvidences);
if( ! pCPD->IsFactorsDistribFunEqual(pBNet1->GetFactor(1), eps) )
{
ret = TRS_FAIL;
}
for( i = 0; i < nEvidences; i++)
{
delete evidences[i];
}
delete pLearn;
delete pBNet1;
delete pBNet;
return trsResult( ret, ret == TRS_OK ? "No errors" :
"Bad test on SetStatistics");
}
void cfactor_setCounts(CFactor &fac, vector<map<size_t, int> > & counts) {
fac.counts() = counts;
}
int main()
{
PNL_USING
//we create very small model to start inference on it
// the model is from Kevin Murphy's BNT\examples\static\belprop_polytree_gaussain
/*
Do the example from Satnam Alag's PhD thesis, UCB ME dept 1996 p46
Make the following polytree, where all arcs point down
0 1
\ /
2
/ \
3 4
*/
int i;
//create this model
int nnodes = 5;
int numnt = 2;
CNodeType *nodeTypes = new CNodeType[numnt];
nodeTypes[0] = CNodeType(0,2);
nodeTypes[1] = CNodeType(0,1);
intVector nodeAssociation = intVector(nnodes,0);
nodeAssociation[1] = 1;
nodeAssociation[3] = 1;
int nbs0[] = { 2 };
int nbs1[] = { 2 };
int nbs2[] = { 0, 1, 3, 4 };
int nbs3[] = { 2 };
int nbs4[] = { 2 };
int *nbrs[] = { nbs0, nbs1, nbs2, nbs3, nbs4 };
int numNeighb[] = {1, 1, 4, 1, 1};
ENeighborType ori0[] = { ntChild };
ENeighborType ori1[] = { ntChild };
ENeighborType ori2[] = { ntParent, ntParent, ntChild, ntChild };
ENeighborType ori3[] = { ntParent };
ENeighborType ori4[] = { ntParent };
ENeighborType *orient[] = { ori0, ori1, ori2, ori3, ori4 };
CGraph *pGraph;
pGraph = CGraph::Create(nnodes, numNeighb, nbrs, orient);
CBNet *pBNet;
pBNet = CBNet::Create( nnodes, numnt, nodeTypes, &nodeAssociation.front(), pGraph );
//Allocation space for all factors of the model
pBNet->AllocFactors();
for( i = 0; i < nnodes; i++ )
{
//Allocation space for all matrices of CPD
pBNet->AllocFactor(i);
}
//now we need to create data for CPDs - we'll create matrices
CFactor *pCPD;
floatVector smData = floatVector(2,0.0f);
floatVector bigData = floatVector(4,1.0f);
intVector ranges = intVector(2, 1);
ranges[0] = 2;
smData[0] = 1.0f;
CNumericDenseMatrix<float> *mean0 = CNumericDenseMatrix<float>::
Create( 2, &ranges.front(), &smData.front());
bigData[0] = 4.0f;
bigData[3] = 4.0f;
ranges[1] = 2;
CNumericDenseMatrix<float> *cov0 = CNumericDenseMatrix<float>::
Create( 2, &ranges.front(), &bigData.front());
pCPD = pBNet->GetFactor(0);
pCPD->AttachMatrix(mean0, matMean);
pCPD->AttachMatrix(cov0, matCovariance);
ranges[0] = 1;
ranges[1] = 1;
float val = 1.0f;
CNumericDenseMatrix<float> *mean1 = CNumericDenseMatrix<float>::
Create( 2, &ranges.front(), &val );
CNumericDenseMatrix<float> *cov1 = CNumericDenseMatrix<float>::
Create( 2, &ranges.front(), &val );
pCPD = pBNet->GetFactor(1);
pCPD->AttachMatrix(mean1, matMean);
pCPD->AttachMatrix(cov1, matCovariance);
smData[0] = 0.0f;
smData[1] = 0.0f;
ranges[0] = 2;
CNumericDenseMatrix<float> *mean2 = CNumericDenseMatrix<float>::
Create(2, &ranges.front(), &smData.front());
smData[0] = 2.0f;
smData[1] = 1.0f;
CNumericDenseMatrix<float> *w21 = CNumericDenseMatrix<float>::
Create(2, &ranges.front(), &smData.front());
bigData[0] = 2.0f;
bigData[1] = 1.0f;
bigData[2] = 1.0f;
bigData[3] = 1.0f;
ranges[1] = 2;
CNumericDenseMatrix<float> *cov2 = CNumericDenseMatrix<float>::
Create(2, &ranges.front(), &bigData.front());
bigData[0] = 1.0f;
bigData[1] = 2.0f;
bigData[2] = 1.0f;
bigData[3] = 0.0f;
CNumericDenseMatrix<float> *w20 = CNumericDenseMatrix<float>::
Create(2, &ranges.front(), &bigData.front());
pCPD = pBNet->GetFactor(2);
pCPD->AttachMatrix( mean2, matMean );
pCPD->AttachMatrix( cov2, matCovariance );
pCPD->AttachMatrix( w20, matWeights,0 );
pCPD->AttachMatrix( w21, matWeights,1 );
val = 0.0f;
ranges[0] = 1;
ranges[1] = 1;
CNumericDenseMatrix<float> *mean3 = CNumericDenseMatrix<float>::
Create(2, &ranges.front(), &val);
val = 1.0f;
CNumericDenseMatrix<float> *cov3 = CNumericDenseMatrix<float>::
Create(2, &ranges.front(), &val);
ranges[1] = 2;
smData[0] = 1.0f;
smData[1] = 1.0f;
CNumericDenseMatrix<float> *w30 = CNumericDenseMatrix<float>::
Create(2, &ranges.front(), &smData.front());
pCPD = pBNet->GetFactor(3);
pCPD->AttachMatrix( mean3, matMean );
pCPD->AttachMatrix( cov3, matCovariance );
pCPD->AttachMatrix( w30, matWeights,0 );
ranges[0] = 2;
ranges[1] = 1;
smData[0] = 0.0f;
smData[1] = 0.0f;
CNumericDenseMatrix<float> *mean4 = CNumericDenseMatrix<float>::
Create(2, &ranges.front(), &smData.front());
ranges[1] = 2;
bigData[0] = 1.0f;
bigData[1] = 0.0f;
bigData[2] = 0.0f;
bigData[3] = 1.0f;
CNumericDenseMatrix<float> *cov4 = CNumericDenseMatrix<float>::
Create(2, &ranges.front(), &bigData.front());
bigData[2] = 1.0f;
CNumericDenseMatrix<float> *w40 = CNumericDenseMatrix<float>::
Create(2, &ranges.front(), &bigData.front());
pCPD = pBNet->GetFactor(4);
pCPD->AttachMatrix( mean4, matMean );
pCPD->AttachMatrix( cov4, matCovariance );
pCPD->AttachMatrix( w40, matWeights,0 );
//Generate random evidences for the modes
int nEv = 1000;
pEvidencesVector evid;
pBNet->GenerateSamples( &evid, nEv );
/////////////////////////////////////////////////////////////////////
//Create copy of initial model with random matrices
CGraph *pGraphCopy = CGraph::Copy(pGraph);
CBNet *pLearnBNet = CBNet::CreateWithRandomMatrices(pGraphCopy, pBNet->GetModelDomain() );
// Creating learning process
CEMLearningEngine *pLearn = CEMLearningEngine::Create(pLearnBNet);
pLearn->SetData(nEv, &evid.front());
pLearn->Learn();
CNumericDenseMatrix<float> *pMatrix;
int length = 0;
const float *output;
///////////////////////////////////////////////////////////////////////
std::cout<<" results of learning (number of evidences = "<<nEv<<std::endl;
for (i = 0; i < nnodes; i++ )
{
int j;
std::cout<<"\n matrix mean for node "<<i;
std::cout<<"\n initial BNet \n";
pMatrix = static_cast<CNumericDenseMatrix<float>*>
(pBNet->GetFactor(i)->GetMatrix(matMean));
pMatrix->GetRawData(&length, &output);
for ( j = 0; j < length; j++ )
{
std::cout<<" "<<output[j];
}
std::cout<<"\n BNet with random matrices after learning \n ";
pMatrix = static_cast<CNumericDenseMatrix<float>*>
(pLearnBNet->GetFactor(i)->GetMatrix(matMean));
pMatrix->GetRawData(&length, &output);
for ( j = 0; j < length; j++)
{
std::cout<<" "<<output[j];
}
std::cout<<"\n \n matrix covariance for node "<<i<<'\n';
std::cout<<"\n initial BNet \n";
pMatrix = static_cast<CNumericDenseMatrix<float>*>
(pBNet->GetFactor(i)->GetMatrix(matCovariance));
pMatrix->GetRawData(&length, &output);
for (j = 0; j < length; j++ )
{
std::cout<<" "<<output[j];
}
std::cout<<"\n BNet with random matrices after learning \n ";
pMatrix = static_cast<CNumericDenseMatrix<float>*>
(pLearnBNet->GetFactor(i)->GetMatrix(matCovariance));
pMatrix->GetRawData(&length, &output);
for ( j = 0; j < length; j++ )
{
std::cout<<" "<<output[j];
}
std::cout<<"\n ___________________________\n";
}
for( i = 0; i < nEv; i++)
{
delete evid[i];
}
delete pLearn;
delete pLearnBNet;
delete pBNet;
return 1;
}
void CSoftMaxCPD::CreateMeanAndCovMatrixForNode(int Node, const CEvidence* pEvidence,
const CBNet *pBNet, floatVector &Mean,
C2DNumericDenseMatrix<float>**CovContParents) const
{
int i, j, k;
int *multiindex = new int [2];
intVector contParents;
pBNet->GetContinuousParents(Node, &contParents);
intVector lineSizes;
lineSizes.assign(2, contParents.size());
floatVector zerodata;
zerodata.assign(contParents.size()*contParents.size(), 0.0f);
*CovContParents =
C2DNumericDenseMatrix<float>::Create( &lineSizes.front(), &zerodata.front() );
for (i = 0; i < contParents.size(); i++ )
{
for (j = 0; j < contParents.size(); j++ )
{
multiindex[0] = i;
multiindex[1] = j;
(*CovContParents)->SetElementByIndexes(0.0f, multiindex);
}
}
intVector pObsNodes;
pConstValueVector pObsValues;
pConstNodeTypeVector pNodeTypes;
pEvidence->GetObsNodesWithValues(&pObsNodes, &pObsValues, &pNodeTypes);
Mean.resize(0);
for ( i = 0; i < contParents.size(); i++)
{
int CurNode = contParents[i];
CFactor *param;
param = pBNet->GetFactor(CurNode);
intVector parentOut;
intVector DiscrParents;
pBNet->GetDiscreteParents(CurNode, &DiscrParents);
int *parentComb = new int [DiscrParents.size()];
for ( j = 0; j < DiscrParents.size(); j++)
{
for ( k = 0; k < pObsNodes.size(); k++)
{
if (DiscrParents[j] == pObsNodes[k])
{
parentComb[j] = pObsValues[k]->GetInt();
break;
}
}
}
const float *data;
int datasize;
static_cast<CDenseMatrix<float>*>(param->GetMatrix(matMean, -1, parentComb))->
GetRawData(&datasize, &data);
Mean.push_back(data[0]);
static_cast<CDenseMatrix<float>*>(param->GetMatrix(matCovariance, -1, parentComb))->
GetRawData(&datasize, &data);
multiindex[0] = i;
multiindex[1] = i;
(*CovContParents)->SetElementByIndexes(data[0], multiindex);
delete [] parentComb;
}
delete [] multiindex;
}
