25 #include <queso/InvLogitGaussianJointPdf.h>
26 #include <queso/GslVector.h>
27 #include <queso/GslMatrix.h>
32 template<
class V,
class M>
36 const V& lawExpVector,
37 const V& lawVarVector)
39 BaseJointPdf<V,M>(((std::string)(prefix)+
"invlogit_gau").c_str(),
41 m_lawExpVector(new V(lawExpVector)),
42 m_lawVarVector(new V(lawVarVector)),
43 m_diagonalCovMatrix(true),
44 m_lawCovMatrix(m_domainSet.vectorSpace().newDiagMatrix(lawVarVector)),
45 m_domainBoxSubset(domainBoxSubset)
50 template<
class V,
class M>
54 const V& lawExpVector,
55 const M& lawCovMatrix)
57 BaseJointPdf<V,M>(((std::string)(prefix)+
"invlogit_gau").c_str(),
59 m_lawExpVector(new V(lawExpVector)),
60 m_lawVarVector(domainBoxSubset.vectorSpace().newVector(INFINITY)),
61 m_diagonalCovMatrix(false),
62 m_lawCovMatrix(new M(lawCovMatrix)),
63 m_domainBoxSubset(domainBoxSubset)
67 template<
class V,
class M>
70 delete m_lawCovMatrix;
71 delete m_lawVarVector;
72 delete m_lawExpVector;
75 template <
class V,
class M>
79 return *m_lawExpVector;
82 template <
class V,
class M>
86 return *m_lawVarVector;
89 template<
class V,
class M>
92 const V& domainVector,
93 const V* domainDirection,
96 V* hessianEffect)
const
100 returnValue = std::exp(this->lnValue(domainVector,domainDirection,gradVector,hessianMatrix,hessianEffect));
105 template<
class V,
class M>
108 const V& domainVector,
109 const V* domainDirection,
112 V* hessianEffect)
const
115 double lnDeterminant = 0.0;
116 V transformedDomainVector(domainVector);
118 V min_domain_bounds(this->m_domainBoxSubset.minValues());
119 V max_domain_bounds(this->m_domainBoxSubset.maxValues());
121 double lnjacobian = 0.0;
122 for (
unsigned int i = 0; i < domainVector.sizeLocal(); i++) {
123 double min_val = min_domain_bounds[i];
124 double max_val = max_domain_bounds[i];
126 if (boost::math::isfinite(min_val) &&
127 boost::math::isfinite(max_val)) {
129 if (domainVector[i] == min_val || domainVector[i] == max_val) {
135 transformedDomainVector[i] = std::log(domainVector[i] - min_val) -
136 std::log(max_val - domainVector[i]);
138 lnjacobian += std::log(max_val - min_val) -
139 std::log(domainVector[i] - min_val) -
140 std::log(max_val - domainVector[i]);
142 else if (boost::math::isfinite(min_val) &&
143 !boost::math::isfinite(max_val)) {
145 if (domainVector[i] == min_val) {
152 transformedDomainVector[i] = std::log(domainVector[i] - min_val);
154 lnjacobian += -std::log(domainVector[i] - min_val);
156 else if (!boost::math::isfinite(min_val) &&
157 boost::math::isfinite(max_val)) {
159 if (domainVector[i] == max_val) {
166 transformedDomainVector[i] = -std::log(max_val - domainVector[i]);
168 lnjacobian += -std::log(max_val - domainVector[i]);
172 transformedDomainVector[i] = domainVector[i];
176 V diffVec(transformedDomainVector - this->lawExpVector());
177 if (m_diagonalCovMatrix) {
178 returnValue = ((diffVec * diffVec) /
179 this->lawVarVector()).sumOfComponents();
180 if (m_normalizationStyle == 0) {
181 unsigned int iMax = this->lawVarVector().sizeLocal();
182 for (
unsigned int i = 0; i < iMax; ++i) {
183 lnDeterminant += log(this->lawVarVector()[i]);
188 V tmpVec = this->m_lawCovMatrix->invertMultiply(diffVec);
189 returnValue = (diffVec * tmpVec).sumOfComponents();
190 if (m_normalizationStyle == 0) {
191 lnDeterminant = this->m_lawCovMatrix->lnDeterminant();
194 if (m_normalizationStyle == 0) {
195 returnValue += ((double) this->lawVarVector().sizeLocal()) * log(2 * M_PI);
196 returnValue += lnDeterminant;
199 returnValue += m_logOfNormalizationFactor;
200 returnValue += lnjacobian;
205 template<
class V,
class M>
211 if ((m_env.subDisplayFile()) && (m_env.displayVerbosity() >= 2)) {
212 *m_env.subDisplayFile() <<
"Entering GaussianJointPdf<V,M>::computeLogOfNormalizationFactor()"
216 if ((m_env.subDisplayFile()) && (m_env.displayVerbosity() >= 2)) {
217 *m_env.subDisplayFile() <<
"Leaving GaussianJointPdf<V,M>::computeLogOfNormalizationFactor()"
218 <<
", m_logOfNormalizationFactor = " << m_logOfNormalizationFactor
225 template<
class V,
class M>
230 delete m_lawExpVector;
231 m_lawExpVector =
new V(newLawExpVector);
234 template<
class V,
class M>
239 delete m_lawCovMatrix;
240 m_lawCovMatrix =
new M(newLawCovMatrix);
243 template<
class V,
class M>
247 return *m_lawCovMatrix;
const V & lawExpVector() const
Access to the vector of mean values of the Gaussian (not transformed) and private attribute: m_lawExp...
~InvLogitGaussianJointPdf()
Destructor.
void updateLawCovMatrix(const M &newLawCovMatrix)
Updates the lower triangular matrix from Cholesky decomposition of the covariance matrix to the new v...
A templated (base) class for handling joint PDFs.
double commonComputeLogOfNormalizationFactor(unsigned int numSamples, bool updateFactorInternally) const
Common method (to the derived classes) to compute the logarithm of the normalization factor...
const V & lawVarVector() const
Access to the vector of variance values and private attribute: m_lawVarVector.
double lnValue(const V &domainVector, const V *domainDirection, V *gradVector, M *hessianMatrix, V *hessianEffect) const
Logarithm of the value of the (transformed) Gaussian PDF (scalar function).
A class for handling hybrid (transformed) Gaussians with bounds.
double actualValue(const V &domainVector, const V *domainDirection, V *gradVector, M *hessianMatrix, V *hessianEffect) const
Actual value of the (transformed) Gaussian PDF.
InvLogitGaussianJointPdf(const char *prefix, const BoxSubset< V, M > &domainBoxSubset, const V &lawExpVector, const V &lawVarVector)
Constructor.
const M & lawCovMatrix() const
Returns the covariance matrix; access to protected attribute m_lawCovMatrix.
Class representing a subset of a vector space shaped like a hypercube.
void updateLawExpVector(const V &newLawExpVector)
Updates the mean of the Gaussian (not transformed) with the new value newLawExpVector.
double computeLogOfNormalizationFactor(unsigned int numSamples, bool updateFactorInternally) const
Computes the logarithm of the normalization factor.