Openturns

=== Library ===

==== Major changes ====
* Added OPT++ solvers
* Improved a lot the performance of all the Rosenblatt related computations
* Added elementary calibration capabilities
* Added CMinpack, Ceres Solver least-squares solvers

==== New classes ====
* ParametricPointToFieldFunction
* LinearModelResult, LinearModelAlgorithm, LinearModelAnalysis
* OPTpp
* NearestPointProblem, LeastSquaresProblem
* CMinpack
* Ceres
* DiscreteMarkovChain
* CalibrationAlgorithm, CalibrationResult, GaussianLinearCalibration, LinearLeastSquaresCalibration
* NonLinearLeastSquaresCalibration, GaussianNonLinearCalibration
* Hypergeometric

==== API changes ====
* Removed deprecated FunctionalChaosRandomVector::getSobol* methods
* Removed deprecated UserDefinedCovarianceModel constructor based on a Collection<CovarianceMatrix>
* Removed deprecated FieldFunction,FieldToPointFunction::getSpatialDimension
* Removed deprecated LinearModelRSquared, LinearModelAdjustedRSquared
* Deprecated LinearModel, LinearModelFactory
* Moved HypothesisTest::(Partial|Full)Regression to LinearModelTest
* Deprecated OptimizationProblem(levelFunction, levelValue)
* Deprecated (Linear|Quadratic)(LeastSquares|Taylor)::getResponseSurface
* FittingTest::BestModelBIC returns the Distribution and the BIC value
* Deprecated VisualTest::DrawEmpiricalCDF,DrawHistogram,DrawClouds
* Add statistic attribute and accessor to the TestResult class
* Deprecated Point::getDescription,setDescription

=== Python module ===

=== Miscellaneous ===
* Changed bugtracker to GitHub issues (https://github.com/openturns/openturns/issues)
* Dropped rot dependency

=== Bug fixes ===
* 289 (Move the LinearModelFactory class from the Base namespace to the Uncertainty namespace)
* 579 (There is no example of a typical "Central Tendency" study)
* 839 (more optim solvers details)
* 931 (The PostAnalyticalImportanceSampling and PostAnalyticalControlledImportanceSampling are too briefly documented)
* 979 (Viewer does not handle BarPlot's fillStyle)
* 977 (HypothesisTest::ChiSquared is bogus)
* 980 (CorrelationAnalysis_SRC scales the coefficients so that they sum to 1)
* 981 (HypothesisTest_{Full, Partial} regression should be moved to LinearModelTest)
* 982 (LinearModelTest::LinearModelDurbinWatson with several factors)
* 983 (Bogus Normal parameter distribution)
* 989 (Fixed the documentation of BIC according to the code)
* 995 (Pip does not recognize conda install of openturns)
* 1000 (SaltelliSensitivityAlgorithm with LowDiscrepancyExperiment produces wrong results)
* 1005 (The stopping criteria of SobolSimulationAlgorithm is weird)
* 1024 (The unary operator is undefined for Normal distribution)
* 1025 (Empty throw in PythonWrappingFunctions)
* 1028 (Build fails with clang-60: token is not a valid binary operator in a preprocessor subexpression)
* 1031 (The doc for LogNormal and LogNormalMuSigma is confusing )
* 1035 (ProbabilitySimulationResult does not provide the distribution of the probability)
* 1036 (Doc error for KrigingAlgorithm method getReducedLogLikelihoodFunction())
* 1043 (The DrawSobolIndices doc is wrong)
* 1045 (LowDiscrepancyExperiment/ComposedCopula correlation across blocks ?)
* 1051 (The default value of computeSecondOrder in SobolIndicesExperiment should be False)
* 1054 (Optimization algorithms ignore MaxEvaluationNumber parameter)
* 1057 (FittingTest.BestModelBIC to return bic value)
* 1058 (GEV problem)
* 1064 (The summary of a FunctionalChaosSobolIndices fails with more than 14 dimensions)
* 1071 (Event cannot be created from a RandomVector)
* 1078 (KrigingAlgorithm documentation still references "inputTransformation")
* 1080 (Brent, Secant, Bissection documentation)
* 1085 (GPD documentation)
* 1090 (PythonFunction can make OT crash)
* 1092 (The parameterGradient of a ParametricFunction can loss accuracy)
* 1099 (PythonDistribution does not work with ConditionalDistribution)
* 1111 (Function::draw() does not take the scale parameter properly into account)
* 1112 (Several bugs in Multinomial)
* 1119 (The parameter description is not taken into account in CalibrationResult)
* 1129 (Segmentation fault with RandomMixture)
* 1139 (CopulaImplementation should derive from DistributionImplementation)
* 1143 (The ResourceMap is not correctly formatted in help pages)
* 1148 (Fails to find cminpack)
* 1155 (The description of Points can be set, but get is empty)

=== Library ===

==== Major changes ====
* FieldFunction and co knows its input/output meshes before evaluation
* SobolSequence has been extended from maximum dimension 40 to 1111
* Parametrized statistical tests by first kind risk instead of 1-risk
* FittingTest now compute a correct p-value in Kolmogorov tests even if the parameters are estimated from the tested sample
* Completed documentation migration with process content

==== New classes ====
* ProbabilitySimulationResult
* ExpectationSimulationAlgorithm, ExpectationSimulationResult
* ExtremeValueCopula
* ChaospyDistribution
* SobolSimulationAlgorithm, SobolSimulationResult
* FractionalBrownianMotionModel

==== API changes ====
* Removed deprecated MonteCarlo, ImportanceSampling, QuasiMonteCarlo, RandomizedQuasiMonteCarlo, RandomizedLHS classes
* Removed deprecated Field::getDimension, getSpatialDimension, getSpatialMean, getTemporalMean methods
* Remove deprecated Process::getDimension, getSpatialDimension methods
* Removed deprecated CovarianceModel::getDimension, getSpatialDimension, getSpatialCorrelation, setSpatialCorrelation methods
* Removed deprecated SpectralModel::getDimension, getSpatialDimension, getSpatialCorrelation methods
* Removed deprecated TruncatedDistribution single bound accessors
* Removed deprecated Function constructors
* Removed deprecated Sample,SampleImplementation::operator*(SquareMatrix) and operator/(SquareMatrix) (and in-place operators)
* Removed deprecated SampleImplementation::scale(SquareMatrix)
* Removed deprecated Domain(a, b) constructor
* Removed deprecated Domain,DomainImplementation::numericallyContains, isEmpty, isNumericallyEmpty, getVolume, getNumericalVolume, computeVolume
* Removed deprecated Domain,DomainImplementation::getLowerBound, getUpperBound methods
* Removed deprecated Interval,Mesh::computeVolume
* Removed deprecated SobolIndicesAlgorithm::[sg]etBootstrapConfidenceLevel
* Removed deprecated Mesh::getVerticesToSimplicesMap,computeSimplexVolume
* Removed deprecated Evaluation,EvaluationImplementation,EvaluationProxy,Function,FunctionImplementation,ParametricEvaluation::operator(inP,parameter)
* Removed deprecated Gradient,GradientImplementation,ParametricGradient::gradient(inP,parameter)
* Removed deprecated Hessian,HessianImplementation,ParametricHessian::operator(inP,parameter)
* Removed ExponentialCauchy, SecondOrderModel, SecondOrderImplementation
* LowDiscrepancyExperiment to work with dependent distributions
* Deprecated UserDefinedCovarianceModel constructor based on a Collection<CovarianceMatrix>
* Deprecated FieldFunction,FieldToPointFunction::getSpatialDimension
* Removed VertexFunction class
* Deprecated LinearModelRSquared, LinearModelAdjustedRSquared
* Added ResourceMap::GetType to access the type of a given key.
* Extended OrthogonalBasis::build() to multi-indices (see ticket 967)
* Deprecated specific RandomVector constructors

=== Python module ===

=== Miscellaneous ===
* CMake >=2.8.8 is required to build OpenTURNS
* The format of openturns.conf has changed to make a distinction between types of keys

=== Bug fixes ===
* 660 (Strange behavior of KernelSmoothing)
* 684 (Strange behavior of Student distribution in multidimensional case)
* 778 (DirectionalSampling is unstable)
* 915 (Distribution.computeQuantile properties)
* 949 (LowDiscrepancyExperiment for distributions having dependent copula)
* 952 (Limitation on distribution type for polynomial chaos?)
* 953 (RandomMixture can fail with truncated distribution)
* 954 (CompositeProcess::getFuture() broken)
* 955 (StudentFactory does not estimate the standard deviation)
* 957 (RandomMixture::getDispersionIndicator() takes a long time)
* 958 (The Sobol' indices plot is wrong for chaos)
* 959 (The PolygonArray and Polygon classes poorly manage the colors)
* 960 (Polygon sometimes make OT crash)
* 961 (The PolygonArray class seem to ignore the legends)
* 962 (The doc of MetaModelValidation.computePredictivityFactor is wrong)
* 963 (VisualTest_DrawHistogram sometimes has overlapping X labels)
* 964 (Study does not load LogNormalMuSigma variables from XML)
* 965 (A RandomVector from a RandomVector can make OT crash)
* 968 (Empty legend make OT crash)
* 970 (Composing gaussian copulas can crash the chaos)
* 972 (FittingTest::ChiSquared slow and buggy)
* 974 (Default constructor of the TruncatedDistribution)
* 975 (EmpiricalBernsteinCopula is a copula as sample is truncated)

=== Library ===

==== Major changes ====

==== New classes ====
* FrechetFactory
* Fehlberg
* TranslationFunction
* DomainComplement, DomainIntersection, DomainUnion, DomainDisjunctiveUnion, DomainDifference
* SmoothedUniform
* NearestNeighbourAlgorithm, RegularGridNearestNeighbour, NaiveNearestNeighbour, NearestNeighbour1D
* EnclosingSimplex, EnclosingSimplexImplementation, NaiveEnclosingSimplex, RegularGridEnclosingSimplex, EnclosingSimplexMonotonic1D, BoundingVolumeHierarchy
* IndicesCollection
* SymbolicParserExprTk
* EvaluationProxy
* MemoizeFunction, MemoizeEvaluation
* Evaluation, Gradient, Hessian
* MeshDomain
* NormInfEnumerateFunction
* FunctionalChaosSobolIndices
* P1LagrangeInterpolation
* FilonQuadrature

==== API changes ====
* Removed deprecated NumericalMathFunction class
* Removed deprecated QuadraticNumericalMathFunction class
* Removed deprecated LinearNumericalMathFunction class
* Removed deprecated NumericalSample class
* Removed deprecated NumericalPoint[WithDescription] class
* Removed deprecated NumericalScalarCollection class
* Removed deprecated NumericalComplexCollection class
* Removed deprecated PosteriorRandomVector class
* Removed deprecated ConditionedNormalProcess class
* Removed deprecated ResourceMap::[SG]AsNumericalScalar methods
* Removed deprecated SpecFunc::*NumericalScalar* constants
* Removed deprecated PlatformInfo::GetConfigureCommandLine method
* Removed deprecated Field::getSample method
* Removed deprecated SobolIndicesAlgorithm::Generate method
* Removed deprecated NumericalScalar, NumericalComplex types
* Removed deprecated Function constructors
* CorrelationAnalysis::(Pearson|Spearman)Correlation accepts a multivariate sample and returns a Point
* Deprecated Field::getDimension, getSpatialDimension, getSpatialMean, getTemporalMean
* Deprecated Process::getDimension, getSpatialDimension
* Deprecated CovarianceModel::getDimension, getSpatialDimension, getSpatialCorrelation, setSpatialCorrelation
* Deprecated SecondOrderModel::getDimension, getSpatialDimension
* Deprecated SpectralModel::getDimension, getSpatialDimension, getSpatialCorrelation
* Deprecated TruncatedDistribution single bound accessors in favor of setBounds/getBounds
* Deprecated remaining analytical & database Function ctors
* Mesh constructor no longer builds a KDTree, new method Mesh::setNearestNeighbourAlgorithm must be called explicitly
* Deprecated Domain::numericallyContains, isEmpty, isNumericallyEmpty, getVolume, getNumericalVolume, computeVolume, getLowerBound, getUpperBound
* Add an argument to KarhunenLoeveQuadratureAlgorithm to pass domain bounds
* Add an argument to Mesh::streamToVTKFile() and Mesh::exportToVTKFile() to pass custom simplices
* All KDTree methods are modified
* IndicesCollection is no more an alias to Collection<Indices>, it has its own class and has a fixed size
* Removed BipartiteGraph::add method
* Deprecated usage of _e and _pi constants when defining analytical functions, e_ and pi_ must be used instead.
* Removed Function::enableHistory,disableHistory,isHistoryEnabled,clearHistory,getHistoryInput,getHistoryOutput,getInputPointHistory,getInputParameterHistory
* Removed Function::enableCache,disableCache,isCacheEnabled,getCacheHits,addCacheContent,getCacheInput,getCacheOutput,clearCache
* Added Drawable::getPaletteAsNormalizedRGBA() to get the palette into a native matplotlib format.
* Changed DistributionImplementation::getCopula,getMarginal,getStandardRepresentative,getStandardDistribution to return a Distribution instead of Pointer<DistributionImplementation>
* Changed DistributionFactoryImplementation::build to return a Distribution instead of Pointer<DistributionImplementation>
* Changed ProcessImplementation::getMarginal to return a Process instead of Pointer<ProcessImplementation>
* Changed RandomVector::getAntecedent,getMarginal to return a RandomVector instead of Pointer<RandomVectorImplementation>
* Changed all {get,set}{Evaluation,Gradient,Hessian} methods to work on Evaluation/Gradient/Hessian instead of Pointer
* Deprecated SobolIndicesAlgorithm::setBootstrapConfidenceLevel, getBootstrapConfidenceLevel
* TNC, Cobyla, EGO are parametrized by setMaximumEvaluationNumber instead of setMaximumIterationNumber
* Removed all KDTree services from Mesh,RegularGrid,Field::getNearestVertex, getNearestVertexIndex, etc, client code must use NearestNeighbourAlgorithm if needed
* Changed Mesh to not derive from DomainImplementation
* Deprecated Mesh::computeSimplexVolume, use Mesh::computeSimplicesVolume instead
* Deprecated Mesh,Interval::computeVolume
* Deprecated FunctionalChaosRandomVector::getSobol* methods in favor of FunctionalChaosSensitivity ones
* DeprecatedEvaluation,EvaluationImplementation,EvaluationProxy,Function,FunctionImplementation,ParametricEvaluation::operator()(inP,parameter)
* Deprecated Gradient,GradientImplementation,ParametricGradient::gradient(inP,parameter)
* Deprecated Hessian,HessianImplementation,ParametricHessian::hessian(inP,parameter)

=== Python module ===
* Depend on swig>=2.0.9

=== Miscellaneous ===
* Multivariate TruncatedDistribution
* Asymptotic Sobol' variance estimators

=== Bug fixes ===
* 870 (Problem with TensorizedCovarianceModel with spatial dim > 1)
* 926 (Covariance model active parameter set behavior)
* 928 (Covariance model/ Field/Process properties naming)
* 932 (Optim algo iteration/evaluation number)
* 937 (Small bugs in FunctionalChaosResult::get{Residuals, RelativeErrors})
* 938 (DistributionFactory::buildEstimator to handle Exception)
* 939 (The help of the getMaximumDegreeStrataIndex method is wrong.)
* 941 (Duplicate with SobolIndicesExperiment)
* 944 (Inline plots crash in notebooks if too many points)
* 945 (build fails with cmake 3.11)
* 946 (OT via anaconda, usage of ot.HypothesisTest error: R_EXECUTABLE-NOTFOUND)
* 947 (EmpiricalBernsteinCopula::setCopulaSample() too restrictive)
* 948 (Bug with BayesDistribution)
* 950 (Circular call to getShape() in Copula)

=== Library ===

==== Major changes ====
* It is now possible to define numerical model acting on either Point or Field
to produce either Point or Field.
The Python binding has been extended to allow the user to define such
functions based on either a Python function or a Python class.
All the possible compositions have been implemented.

==== New classes ====
* ProbabilitySimulation
* SobolIndicesExperiment
* VertexFunction
* FieldToPointFunction
* FieldToPointFunctionImplementation
* PythonFieldToPointFunction
* OpenTURNSPythonFieldToPointFunction
* PointToFieldFunction
* PointToFieldFunctionImplementation
* PythonPointToFieldFunction
* OpenTURNSPythonPointToFieldFunction
* KarhunenLoeveLifting
* KarhunenLoeveProjection
* PointToPointEvaluation
* PointToPointConnection
* PointToFieldConnection
* FieldToFieldConnection
* FieldToPointConnection

==== API changes ====
* Removed deprecated LAR class
* Remove deprecated Function::GetValidConstants|GetValidFunctions|GetValidOperators
* Removed deprecated TemporalNormalProcess, SpectralNormalProcess classes
* Removed deprecated GeneralizedLinearModelAlgorithm, GeneralizedLinearModelResult classes
* Removed deprecated DynamicalFunction, SpatialFunction, TemporalFunction classes
* Removed deprecated KarhunenLoeveP1Factory, KarhunenLoeveQuadratureFactory classes
* Removed deprecated GramSchmidtAlgorithm, ChebychevAlgorithm classes
* Removed [gs]etOptimizationSolver methods
* Removed CovarianceModel::compute{AsScalar,StandardRepresentative} overloads
* Deprecated PosteriorRandomVector
* Deprecated MonteCarlo, ImportanceSampling, QuasiMonteCarlo, RandomizedQuasiMonteCarlo, RandomizedLHS classes
* Made MatrixImplementation::isPositiveDefinite const and removed its argument
* Renamed EfficientGlobalOptimization::setAIETradeoff to setAEITradeoff
* Deprecated PlatformInfo::GetConfigureCommandLine.
* Renamed ConditionedNormalProcess to ConditionedGaussianProcess
* Deprecated Field::getSample in favor of getValues
* Deprecated SobolIndicesAlgorithm::Generate in favor of SobolIndicesExperiment.

=== Python module ===

=== Miscellaneous ===
* Changed bounds evaluation in UniformFactory, BetaFactory
* Worked around bug 864 (parallel segfault in BernsteinCopulaFactory)

=== Bug fixes ===
* 890 (Cannot build triangular distribution)
* 891 (Viewer issue with Pairs drawables)
* 895 (Trouble reading CSV files with separators in description)
* 896 (Python iteration in ProcessSample leads to capacity overflow)
* 897 (Bug in Graph::draw with small data)
* 898 (Could not save/load some persistent classes)
* 899 (PythonDistribution copula crash when parallelism is active)
* 902 (NormalGamma constructor builds wrong link function)
* 905 (Bogus MaternModel::setParameter)
* 906 (t_LevelSetMesher_std fails on most non-Intel based chips)
* 907 (Notation)
* 908 (Documentation: change titles)
* 909 (Wrong argument type in the API doc)
* 910 (Graph of a d-dimensionnal distribution)
* 911 (In the Field class, the getSample and getValues methods are duplicate)
* 912 (Wrong description of Histogram constructor parameters)
* 914 (KarhunenLoeveQuadratureAlgorithm crashes for covariance models of dimension>1)
* 917 (Bug in RandomMixture::computeCDF())
* 918 (The class SobolIndicesAlgorithm has a draw method which has no example)
* 919 (Wrong simplification mechanism in MarginalTransformationEvaluation for Exponential distribution)
* 921 (Cannot print a FixedExperiment when built from sample and weight)
* 923 (Fix ExponentialModel::getParameter for diagonal correlation)
* 924 (Probleme with the factory of a Generalized Pareto distribution)
* 927 (Functional chaos is memory hungry)
* 929 (The labels of the sensitivity analysis graphics are poor)
* 930 (The getMean method has a weird behavior on parametrized distribution)

=== Library ===

==== Major changes ====
* Integrate otlhs module
* New function API
* Canonical format low-rank tensor approximation
* EGO global optimization algorithm

==== New classes ====
* SpaceFillingPhiP, SpaceFillingMinDist, SpaceFillingC2
* LinearProfile, GeometricProfile
* MonteCarloLHS, SimulatedAnnealingLHS
* LHSResult
* MultiStart
* MethodOfMomentsFactory
* SymbolicFunction, AggregatedFunction, ComposedFunction, DatabaseFunction, DualLinearCombinationFunction
* LinearCombinationFunction, LinearFunction, QuadraticFunction, ParametricFunction, IndicatorFunction
* DistributionTransformation
* GeneralizedExtremeValue
* UniVariateFunctionFamily, UniVariateFunctionFactory, TensorizedUniVariateFunctionFactory
* MonomialFunction, MonomialFunctionFactory
* KarhunenLoeveSVDAlgorithm
* RankMCovarianceModel
* SparseMethod
* CanonicalTensorEvaluation|Gradient, TensorApproximationAlgorithm|Result
* GaussLegendre
* EfficientGlobalOptimization

==== API changes ====
* Removed deprecated SLSQP, LBFGS and NelderMead classes
* Removed deprecated QuadraticCumul class
* Removed classes UserDefinedPair, HistogramPair
* Removed deprecated method WeightedExperiment::getWeight
* Removed deprecated method DistributionFactory::build(NumericalSample, CovarianceMatrix&)
* Removed deprecated distributions alternative parameters constructors, accessors
* Added a generic implementation of the computeLogPDFGradient() method in the DistributionImplementation class.
* Allow Box to support bounds
* Deprecated LinearNumericalMathFunction in favor of LinearFunction
* Deprecated QuadraticNumericalMathFunction in favor of QuadraticFunction
* Deprecated NumericalMathFunction::GetValidConstants|GetValidFunctions|GetValidOperators
* Renamed ComposedNumericalMathFunction to ComposedFunction
* Renamed LinearNumericalMathFunction to LinearFunction
* Swap covModel and basis arguments in KrigingAlgorithm constructors
* Removed useless keepCholesky argument in KrigingAlgorithm constructors
* Renamed OptimizationSolver to OptimizationAlgorithm
* Renamed TemporalNormalProcess to GaussianProcess
* Renamed SpectralNormalProcess to SpectralGaussianProcess
* Renamed GeneralizedLinearModelAlgorithm to GeneralLinearModelAlgorithm
* Renamed GeneralizedLinearModelResult to GeneralLinearModelResult
* Renamed DynamicalFunction to FieldFunction
* Renamed SpatialFunction to ValueFunction
* Renamed TemporalFunction to VertexValueFunction
* Deprecated [gs]etOptimizationSolver methods
* Renamed ProductNumericalMathFunction to ProductFunction
* Deprecated KarhunenLoeveP1Factory, KarhunenLoeveQuadratureFactory
* Deprecated GramSchmidtAlgorithm, ChebychevAlgorithm
* Added getSobolAggregatedIndices() to FunctionalChaosRandomVector
* Added computeWeight() to Mesh
* Deprecated NumericalMathFunction ctors
* Deprecated NumericalMathFunction for Function
* Deprecated NumericalSample for Sample
* Deprecated NumericalPoint[WithDescription] for Point[WithDescription]
* Deprecated ResourceMap::[SG]AsNumericalScalar for [SG]AsScalar
* Deprecated SpecFunc::*NumericalScalar*
* Deprecated NumericalScalar for Scalar
* Deprecated NumericalComplex for Complex
* Deprecated DistributionImplementation::getGaussNodesAndWeights

=== Python module ===

=== Miscellaneous ===

=== Bug fixes ===
* 351 (FORM is it possible to hav a ".setMaximumNumberOfEvaluations")
* 729 (KDTree & save)
* 774 (Exact limits of a normal distribution with unknown mean and variance)
* 866 (Check the parameter estimate for the kriging model)
* 869 (The ProductNumericalMathFunction class has no example)
* 871 (GeneralizedExponential P parameter : int or float ?)
* 872 (Cannot draw a Text drawable using R)
* 874 (Compatibility between a distribution factory and an alternate parametrization not checked)
* 875 (TruncatedNormalFactory randomly crashes)
* 876 (Bad time grid in StationaryCovarianceModelFactory::build)
* 877 (centered whitenoise limitation)
* 878 (Viewer does not take into account labels in Contour)
* 879 (Incomplete arguments in FunctionalChaosRandomVector docstrings)
* 882 (RandomMixture segfaults with Dirac)
* 883 (VisualTest.DrawHistogram should rely on Histogram.drawPDF)
* 886 (Bogus RandomMixture::getSupport)
* 887 (Bogus PDF evaluation in RandomMixture with mix of continuous/discrete variables)
* 888 (Bogus RandomMixture::getSample)

=== Library ===

==== Major changes ====
* Changed the default orthonormalization algorithm of StandardDistributionPolynomialFactory from GramSchmidtAlgorithm to AdaptiveStieltjesAlgorithm
* New api for sensitivity analysis
* New methods to compute confidence regions in Distribution

==== New classes ====
* SubsetSampling
* AdaptiveDirectionalSampling
* KarhunenLoeveQuadratureFactory
* SobolIndicesAlgorithm
* SaltelliSensitivityAlgorithm
* MartinezSensitivityAlgorithm
* JansenSensitivityAlgorithm
* MauntzKucherenkoSensitivityAlgorithm
* SoizeGhanemFactory
* LevelSetMesher
* HistogramPolynomialFactory
* ChebychevFactory
* FourierSeriesFactory, HaarWaveletFactory
* OrthogonalProductFunctionFactory

==== API changes ====
* Removed deprecated (AbdoRackwitz|Cobyla|SQP|TNC)SpecificParameters classes
* Removed AbdoRackwitz|Cobyla|SQP::[gs]etLevelFunction|[gs]etLevelValue
* Removed deprecated OptimizationSolver::setMaximumIterationsNumber
* Removed deprecated method Distribution::setParametersCollection(NP)
* Removed deprecated PersistentFactory string constructor
* Deprecated QuadraticCumul class in favor of TaylorExpansionMoments
* Renamed __contains__ to contains
* Modified NumericalMathFunction::[sg]etParameter to operate on NumericalPoint instead NumericalPointWithDescription
* Add NumericalMathFunction::[sg]etParameterDescription to access the parameter description
* Deprecated classes UserDefinedPair, HistogramPair
* Removed SensitivityAnalysis class
* Deprecated SLSQP, LBFGS and NelderMead classes in favor of NLopt class
* Deprecated LAR in favor of LARS
* Deprecated DistributionFactory::build(NumericalSample, CovarianceMatrix&)
* Deprecated distributions alternative parameters constructors, accessors
* Swap SpectralModel scale & amplitude parameters: CauchyModel, ExponentialCauchy

=== Python module ===
* Added the possibility to distribute PythonFunction calls with multiprocessing

=== Miscellaneous ===
* Improved the computeCDF() method of Normal
* Added the computeMinimumVolumeInterval(), computeBilateralConfidenceInterval(), computeUnilateralConfidenceInterval() and computeMinimumVolumeLevelSet() methods to compute several kind of confidence regions in Distribution
* Added HarrisonMcCabe, BreuschPagan and DurbinWatson tests to test homoskedasticity, autocorrelation of linear regression residuals
* Added two samples Kolmogorov test
* Improved the speed of many algorithms based on method binding
* Added more options to control LHSExperiment and LowDiscrepancyExperiment
* Improved the IntervalMesher class: now it takes into account the diamond flag
* Shortened ResourceMap keys to not contain 'Implementation'
* Improved the performance of Classifier/MixtureClassifier/ExpertMixture

=== Bug fixes ===
* 535 (parallel-threads option cannot be changed at runtime with TBB)
* 565 (The SensitivityAnalysis class manages only one single output.)
* 604 (Bug concerning the NonCentralStudent distribution)
* 698 (KernelSmoothing() as a factory)
* 786 (Bug in sensitivity analysis)
* 802 (Python issue with ComplexMatrix::solveLinearSystem)
* 803 (prefix openturns includes)
* 813 (Error when multiplying a Matrix by a SymmetricMatrix)
* 815 (ConditionedNormalProcess test fails randomly)
* 820 (Python distribution fails randomly when computing the PDF over a sample)
* 822 (Incorect Matrix / point operations with cast)
* 824 (Confusing behavior of NumericalSample::sort)
* 828 (ImportFromCSVFile fails on a file created by exportToCSVFile)
* 830 (more optim algos examples)
* 831 (Missing get/setParameter in OpenTURNSPythonFunction)
* 833 (Homogeneity in Covariance Models)
* 837 (TruncatedDistribution::setParameter segfaults)
* 838 (Symmetry of SymmetricMatrix not always enforced)
* 840 (Remove WeightedExperiment::getWeight)
* 841 (Better CovarianceModelCollection in Python)
* 842 (Better ProcessCollection in Python)
* 843 (Remove all the specific isCopula() methods)
* 848 (Inverse Wishart sampling)
* 849 (Ambiguous NumericalSample::computeQuantile)
* 853 (Switch the default for normalize boolean from TRUE to FALSE in ot.GeneralizedLinearModelAlgorithm)
* 854 (InverseWishart.computeLogPDF)
* 861 (document HMatrix classes)