SG++
sgpp::datadriven Namespace Reference

Namespaces

 clusteringmpi
 
 ClusteringOCL
 
 DBMatOfflineFactory
 factories to build the specialization of the DBMatOffline objects.
 
 DBMatOnlineDEFactory
 
 DensityOCLMultiPlatform
 
 MultipleEvalHPX
 
 PiecewiseConstantRegression
 
 streamingBSplineOCL
 
 StreamingModOCLFastMultiPlatform
 
 StreamingModOCLMaskMultiPlatform
 
 StreamingModOCLOpt
 
 StreamingModOCLUnified
 
 StreamingOCLMultiPlatform
 
 x86simple
 

Classes

class  AbstractOperationMultipleEvalSubspace
 
class  AlgorithmAdaBoostBase
 
class  AlgorithmAdaBoostIdentity
 
class  ArffFileSampleProvider
 ArffFileSampleProvider allows reading data in ARFF format into a sgpp::datadriven::Dataset object. More...
 
class  ARFFTools
 Class that provides functionality to read ARFF files. More...
 
struct  AssignBatchNetworkMessage
 Message wrapped in MPI_Packet specifying an order to a worker to train from a batch. More...
 
struct  AssignSystemMatrixUpdateNetworkMessage
 Message wrapped in MPI_Packet specifying an order to a worker to update a class' system matrix decomposition. More...
 
struct  AssignTaskResult
 Used by the MPI Task Scheduler to deliver the result of assigning the requested task. More...
 
class  BatchLearner
 The Batchlearner learns the data provided as input in batches. More...
 
class  ClassificationLearner
 The ClassificationLearner class Solves a classification problem. More...
 
struct  ClassificatorQuality
 struct to encapsulate the classsifiers quality by its characteristic numbers More...
 
class  ConvergenceMonitor
 A monitor to decide if a learning algorithm has converged. More...
 
class  CrossValidation
 Supervised learning with cross validation used to fit a model and quantify accuracy using a sgpp::datadriven::Metric. More...
 
struct  CrossValidationConfiguration
 Set of parameters required to fully configure sgpp::datadriven::CrossValidation objects. More...
 
struct  CrossvalidationConfiguration
 
class  CrossValidationScorerFactory
 Concrete factory to build an instance of sgpp::datadriven::CrossValidation. More...
 
class  CSVFileSampleProvider
 CSVFileSampleProvider allows reading data in CSV format into a sgpp::datadriven::Dataset object. More...
 
class  CSVTools
 Class that provides functionality to read CSV files. More...
 
class  DataBasedRefinementFunctor
 Data based refinement uses data points to find refinement candidates. More...
 
class  DataMiningConfigParser
 
class  Dataset
 
class  DatasetGenerator
 
class  DatasetTools
 
class  DataSource
 DataSource is a high level, easy to use interface for accessing data provided by a all kinds of sgpp::datadriven::SampleProvider. More...
 
class  DataSourceBuilder
 Generate an instance of sgpp::datadriven::DataSource using the Builder Pattern. More...
 
struct  DataSourceConfig
 Configuration structure used for all kinds of SampleProviders including default values. More...
 
class  DataSourceFileTypeParser
 Convenience class to convert strings to sgpp::datadriven::DataSourceFileType and generate string representations for values of sgpp::datadriven::DataSourceFileType. More...
 
class  DataSourceIterator
 Iterator object for walking convenient walking over the batches of a sgpp::datadriven::DataSource. More...
 
class  DBMatDecompMatrixSolver
 
class  DBMatDMSChol
 Class to solve the system of equations with a LL'-decomposed matrix. More...
 
class  DBMatDMSDenseIChol
 Solve the system of equations with a LL'-decomposed matrix where LL' is created by an iterative, incomplete cholesky factorization on a dense matrix. More...
 
class  DBMatDMSOrthoAdapt
 This class solves an (lhs + lambda*I) * alpha = b system of linear equations after the offline and online phases are done. More...
 
class  DBMatOffline
 Class that is used to decompose and store the left-hand-side matrix for the density based classification approach (The classification is divided into two parts: the offline step that does not depend on the actual data and the online step that depends on the data). More...
 
class  DBMatOfflineChol
 DBMatOffline specialization that uses a cholesky factorization on a dense matrix. More...
 
class  DBMatOfflineDenseIChol
 DBMatOfflineChol specialization that uses a parallel, iterative incomplete cholesky factorization on a dense matrix. More...
 
class  DBMatOfflineGE
 DBMatOffline specialization as a base class for all algorithms based on gaussian elimination on a dense matrix. More...
 
class  DBMatOfflineOrthoAdapt
 
class  DBMatOnline
 Class for objects that can be used in the online step of the classification (The classification is divided into two parts: the offline step that does not depend on the actual data and the online step that depends on the data) More...
 
class  DBMatOnlineDE
 Class that stores, generates and manipulates a density function during online phase in on/off learning. More...
 
class  DBMatOnlineDEChol
 
class  DBMatOnlineDEOrthoAdapt
 
struct  DensityEstimationConfiguration
 
class  DensityEstimationMinerFactory
 Concrete Factory that builds an instance of sgpp::datadriven::SparseGridMiner for Density Estimation. More...
 
class  DensityEstimationTypeParser
 
class  DensityEstimator
 
class  DensitySystemMatrix
 Class that implements the virtual class OperationMatrix for the application of classification for the Systemmatrix by using a density function. More...
 
class  DMSystemMatrix
 Class that implements the virtual class base::OperationMatrix for the application of classification for the Systemmatrix. More...
 
class  DMSystemMatrixBase
 Abstract class that defines the virtual class base::OperationMatrix for classification and regression problems. More...
 
class  DMSystemMatrixBaseSP
 Abstract class that defines the virtual class base::OperationMatrix for classification and regression problems (single precision version) More...
 
class  DMWeightMatrix
 Class that implements the virtual class OperationMatrix for the application of classification for the Systemmatrix with weight. More...
 
class  FileSampleDecorator
 FileSampleDecorator provides an interface to provide generic manipulations for various kinds of sgpp::datadriven::FileSampleProvider using the decorator pattern. More...
 
class  FileSampleProvider
 sgpp::datadriven::FileSampleProvider is an specialization of sgpp::datadriven::SampleProvider and provides an interface for all sample providers that get their samples from files. More...
 
class  FitterConfiguration
 General configuration object for fitters. More...
 
class  FitterConfigurationDensityEstimation
 Configuration for fitter scenarios using density estimation. More...
 
class  FitterConfigurationLeastSquares
 Configuration for fitter scenarios using least squares optimization. More...
 
class  FitterTypeParser
 Convenience class to convert strings to sgpp::datadriven::FitterType and generate string representations for values of sgpp::datadriven::FitterType. More...
 
class  Friedman1Generator
 
class  Friedman2Generator
 
class  Friedman3Generator
 
class  GaussianKDE
 
class  GridPointBasedRefinementFunctor
 Grid Point based refinement for classification problems solved by a SG density estimation approach. More...
 
class  GridTypeParser
 
class  Learner
 This class implements standard sparse grid regression with an arbitrary regularization operator. More...
 
class  LearnerBase
 Abstract class that implements a regression/classification learner based on spatial adaptive Sparse Grids. More...
 
class  LearnerBaseSP
 Abstract class that implements a regression/classification learner based on spatial adaptive Sparse Grids. More...
 
class  LearnerDensityBased
 
class  LearnerDensityBasedReg
 Class that implements a regression learner using density estimation on Sparse Grids. More...
 
class  LearnerLeastSquaresIdentity
 This class implements standard sparse grid regression with an Identity matrix as regularization operator. More...
 
class  LearnerPiecewiseConstantSmoothedRegression
 
class  LearnerScenario
 
class  LearnerSGD
 LearnerSGD learns the data using stochastic gradient descent. More...
 
class  LearnerSGDE
 
class  LearnerSGDEConfiguration
 
class  LearnerSGDEOnOff
 LearnerSGDEOnOff learns the data using sparse grid density estimation. More...
 
class  LearnerSGDEOnOffParallel
 LearnerSGDEOnOffParallel learns the data using sparse grid density estimation. More...
 
class  LearnerSVM
 LearnerSVM learns the data using support vector machines and sparse grid kernels. More...
 
struct  LearnerTiming
 strcut to encapsulate the learner's timings during training More...
 
struct  LearnerVectorizedPerformance
 struct that defines return for calculation the performance of a vectorized learner More...
 
class  LearnerVectorizedPerformanceCalculator
 Class that provides functionality in order to determine a LearnerVectorized's performance. More...
 
class  LeastSquaresRegressionMinerFactory
 Concrete Factory that builds an instance of sgpp::datadriven::SparseGridMiner for Least Squares Regression. More...
 
struct  LevelIndexPair
 Pair to hold the level and the index of a grid point in one dimension. More...
 
class  LogDensitySystemMatrix
 Class that implements the virtual class OperationMatrix for the application of classification for the Systemmatrix by using a density function. More...
 
class  MatrixDecompositionTypeParser
 
struct  MergeGridNetworkMessage
 Packet wrapper in MPI_Packet containing segmented data from the alpha vector of the trained system. More...
 
struct  MessageTrackRequest
 Structure that holds data for pending tracking requests that need to be checked against incoming messages. More...
 
class  MetaLearner
 
class  Metric
 We use metrics to quantify approximation quality of a trained model. More...
 
class  MinerFactory
 Abstract factory to build different kinds of Miners based on a configuration which is parsed from a file. More...
 
class  ModelFittingBase
 Base class for arbitrary machine learning models based on adaptive sparse grids. More...
 
class  ModelFittingDensityEstimation
 Fitter object that encapsulates the usage of sparse grid density estimation with identity as regularization. More...
 
class  ModelFittingLeastSquares
 Fitter object that encapsulates the usage of sparse grid based regression with identity as regularization. More...
 
class  MortonOrder
 Class for re-arranging Datasets along a Morton order curve. More...
 
struct  MPI_Packet
 A packet sent over MPI, using a command as a descriptor, and a wrapped package in the payload for data. More...
 
class  MPIMethods
 
class  MPIRequestPool
 
class  MPITaskScheduler
 
class  MSE
 Metric that quantifies the difference between predicted values and actual values in terms of mean squared error (MSE). More...
 
class  MultiGridRefinementFunctor
 Abstract super-class for refinement functors operating on multiple grids. More...
 
class  MultipleClassRefinementFunctor
 Multiple class refinement is based on the zero-crossing based refinement. More...
 
class  MultiSurplusRefinementFunctor
 Wrapper of SurplusRefinementFunctor for multi grid scenarios. More...
 
class  NearestNeighbors
 The NearestNeighbors class. More...
 
class  OperationDensityConditional
 Conditionalize Probability Density Function. More...
 
class  OperationDensityConditionalKDE
 
class  OperationDensityConditionalLinear
 Marginalize Probability Density Function. More...
 
class  OperationDensityMarginalize
 Marginalize Probability Density Function. More...
 
class  OperationDensityMarginalizeKDE
 Marginalize Probability Density Function. More...
 
class  OperationDensityMarginalizeLinear
 Marginalize Probability Density Function. More...
 
class  OperationDensityMargTo1D
 Marginalize Probability Density Function. More...
 
class  OperationDensityMargTo1DLinear
 keep applying marginalize to function until it's reduced to only 1 dimension More...
 
class  OperationDensityRejectionSampling
 Sampling on all dimensions. More...
 
class  OperationDensityRejectionSamplingLinear
 Sampling with rejection sampling method. More...
 
class  OperationDensitySampling
 Sampling on all dimensions. More...
 
class  OperationDensitySampling1D
 Sample 1D Probability Density Function. More...
 
class  OperationDensitySampling1DLinear
 
class  OperationDensitySamplingLinear
 keep applying marginalize to function until it's reduced to only 1 dimension More...
 
class  OperationDotProductLinear
 
class  OperationDotProductModLinear
 
class  OperationInverseRosenblattTransformation
 Sampling on all dimensions. More...
 
class  OperationInverseRosenblattTransformation1DLinear
 
class  OperationInverseRosenblattTransformationKDE
 Do inverse transformation in all dimensions. More...
 
class  OperationInverseRosenblattTransformationLinear
 keep applying marginalize to function until it's reduced to only 1 dimension More...
 
class  OperationMultiEvalCuda
 OperationMultipleEval for polynomial basis functions (grad >= 2) using CUDA on grids without boundary nodes. More...
 
class  OperationMultiEvalHPX
 This class is a HPX wrapper for other MultiEval-operations that uses a very simple master-slave distributed processing. More...
 
class  OperationMultiEvalModMaskStreaming
 
class  OperationMultiEvalMPI
 This class is a MPI wrapper for other MultiEval-operations that uses a very simple master-slave MPI parallelization. More...
 
class  OperationMultiEvalStreaming
 
class  OperationMultiEvalStreamingBSplineOCL
 
class  OperationMultiEvalStreamingModOCLFastMultiPlatform
 
class  OperationMultiEvalStreamingModOCLMaskMultiPlatform
 
class  OperationMultiEvalStreamingModOCLOpt
 
class  OperationMultiEvalStreamingModOCLUnified
 
class  OperationMultipleEvalConfiguration
 
class  OperationMultipleEvalSubspaceCombined
 Multiple evaluation operation that uses the subspace structure to save work compared to the naive or streaming variants. More...
 
class  OperationMultipleEvalSubspaceSimple
 Multiple evaluation operation that uses the subspace structure to save work compared to the naive or streaming variants. More...
 
class  OperationPiecewiseConstantRegression
 
class  OperationRegularizationDiagonal
 Implementation of the application of a diagonal matrix to a DataVector for regularization. More...
 
class  OperationRegularizationDiagonalLinearBoundary
 Implementation of the application of a diagonal matrix to a DataVector for regularization. More...
 
class  OperationRosenblattTransformation
 Sampling on all dimensions. More...
 
class  OperationRosenblattTransformation1DLinear
 
class  OperationRosenblattTransformationKDE
 Do transformation in all dimensions. More...
 
class  OperationRosenblattTransformationLinear
 keep applying marginalize to function until it's reduced to only 1 dimension More...
 
class  OperationTest
 Operation the tests the function that is applied the current Sparse sgpp::base::Grid at a given point. More...
 
class  OperationTestLinear
 This class implements OperationTest for a grids with linear basis ansatzfunctions without boundaries. More...
 
class  OperationTestLinearBoundary
 This class implements OperationTest for a grids with linear basis ansatzfunctions with boundaries. More...
 
class  OperationTestLinearStretched
 This class implements OperationTest for a grids with linearstretched basis ansatzfunctions without boundaries. More...
 
class  OperationTestLinearStretchedBoundary
 This class implements OperationTest for a grids with linear basis ansatzfunctions with boundaries. More...
 
class  OperationTestModBspline
 This class implements OperationTest for a grids with modified bspline basis functions with a certain degree. More...
 
class  OperationTestModLinear
 This class implements sgpp::base::OperationEval for a grids with mod linear basis ansatzfunctions with. More...
 
class  OperationTestModPoly
 This class implements OperationTest for a grids with mod poly basis ansatzfunctions with. More...
 
class  OperationTestModWavelet
 This class implements OperationTest for a grid with mod wavelet basis ansatzfunctions. More...
 
class  OperationTestPoly
 This class implements OperationTest for a grids with poly basis ansatzfunctions with. More...
 
class  OperationTestPrewavelet
 This class implements OperationTest for a grids with prewavelet basis ansatzfunctions without boundaries. More...
 
class  OperationTransformation1D
 Sample 1D Probability Density Function. More...
 
class  PartitioningTool
 The methods in this class calculate size and offset of a segment for a partition of a domain. More...
 
class  PendingMPIRequest
 
class  PiecewiseConstantSmoothedRegressionMetaLearner
 
class  PiecewiseConstantSmoothedRegressionSystemMatrix
 Class that implements the virtual class OperationMatrix for the application of classification for the Systemmatrix by using a density function. More...
 
class  PrimalDualSVM
 Implementation of a support vector machine in primal formulation which additionally stores support vectors. More...
 
class  RandomShufflingFunctor
 Generate a randomized permutation for a set of indices based on a pseudo random generator. More...
 
class  RefinementHandler
 
struct  RefinementResult
 Structure to hold the grid modifications for a refinement cycle for one class. More...
 
struct  RefinementResultNetworkMessage
 Packet wrapped in an UPDATE_GRID MPI_Packet, containing segmented changes for a specified class. More...
 
struct  RefinementResultSystemMatrixNetworkMessage
 Packet wrapped in a RefinementResultNetwork Message that contains additional information required when updating the system matrix. More...
 
class  RegressionLearner
 The RegressionLearner class Solves a regression problem with continuous target vector. More...
 
struct  RegularizationConfiguration
 
class  RegularizationTypeParser
 
class  RoundRobinScheduler
 
class  SampleProvider
 SampleProvider is an abstraction for object that provide sample data from various sources for example datasets from files (ARFF, CSV) or synthetic datasets generated by sampling functions ( Friedmann datasets). More...
 
class  Scorer
 Base class for supervised learning used to fit a model and quantify accuracy using a sgpp::datadriven::Metric with either testing or cross validation. More...
 
class  ScorerFactory
 Abstract factory to build all kinds of scorers based on a given configuration. More...
 
class  ScorerMetricTypeParser
 Convenience class to convert strings to sgpp::datadriven::ScorerMetricType and generate string representations for values of sgpp::datadriven::ScorerMetricType. More...
 
class  ScorerShufflingTypeParser
 Convenience class to convert strings to sgpp::datadriven::ScorerShufflingType and generate string representations for values of sgpp::datadriven::ScorerShufflingType. More...
 
class  SequentialShufflingFunctor
 Sequential Shuffling does not permute indices at all and thus keeps their order unchanged. More...
 
class  ShufflingFunctor
 A shuffling functor generates a permutation for a set of indices. More...
 
class  SLESolverTypeParser
 Convenience class to convert strings to sgpp::solver::SLESolverType and generate string representations for values of sgpp::solver::SLESolverType. More...
 
class  SortedDataset
 Dataset that can be ordered. Accessing the included DataMatrix might invalidate the order. More...
 
class  SparseGridMiner
 SparseGridMiner models the entire mining process for data mining with sparse grids. More...
 
class  SplittingScorer
 Provide supervised learning functionality. More...
 
class  SplittingScorerFactory
 Concrete factory to build an instance of sgpp::datadriven::SplittingScorer. More...
 
class  StreamingBSplineOCLKernelImpl
 
class  StreamingBSplineOCLKernelSourceBuilder
 
class  StringTokenizer
 
class  SubspaceNodeCombined
 
class  SubspaceNodeSimple
 
class  SystemMatrixLeastSquaresIdentity
 Class that implements the virtual class base::OperationMatrix for the application of classification for the Systemmatrix. More...
 
struct  TestingConfiguration
 Set of parameters required to fully configure sgpp::datadriven::SplittingScorer objects. More...
 
class  TestsetConfiguration
 
class  TunableParameter
 
class  ZeroCrossingRefinementFunctor
 Zero-crossing-based refinement uses zero crossings of the difference PDFS f_1 - f_2 to determine areas of interest for the refinement process. More...
 

Typedefs

typedef std::vector< std::pair< std::unique_ptr< DBMatOnlineDE >, double > > ClassDensityConntainer
 
typedef ClassDensityConntainer ClassDensityContainer
 Type definition to avoid typographical error in original learner. More...
 
typedef std::vector< LevelIndexPairLevelIndexVector
 Vector that holds level index pairs for every dimensions. More...
 

Enumerations

enum  DataSourceFileType { DataSourceFileType::NONE, DataSourceFileType::ARFF, DataSourceFileType::CSV }
 Supported file types for sgpp::datadriven::FileSampleProvider. More...
 
enum  DensityEstimationType { DensityEstimationType::CG, DensityEstimationType::Decomposition }
 
enum  FitterType { FitterType::RegressionLeastSquares, FitterType::DensityEstimation }
 Different fitter scenarios have different default values and support different operations. More...
 
enum  InternalPrecision { InternalPrecision::Float, InternalPrecision::Double }
 
enum  LearnerMode { LearnerMode::LEARN, LearnerMode::LEARNCOMPARE, LearnerMode::LEARNTEST }
 
enum  MatrixDecompositionType {
  MatrixDecompositionType::LU, MatrixDecompositionType::Eigen, MatrixDecompositionType::Chol, MatrixDecompositionType::DenseIchol,
  MatrixDecompositionType::OrthoAdapt
}
 
enum  MPI_COMMAND_ID {
  NULL_COMMAND, UPDATE_GRID, MERGE_GRID, ASSIGN_BATCH,
  COMPUTE_UPDATE_SYSTEM_MATRIX_DECOMPOSITION, SHUTDOWN, WORKER_SHUTDOWN_SUCCESS
}
 Different commands sent over MPI to allow the receiver to identify the message's contents. More...
 
enum  OperationMultipleEvalMPIType { OperationMultipleEvalMPIType::NONE, OperationMultipleEvalMPIType::MASTERSLAVE, OperationMultipleEvalMPIType::HPX }
 
enum  OperationMultipleEvalSubType {
  OperationMultipleEvalSubType::DEFAULT, OperationMultipleEvalSubType::SIMPLE, OperationMultipleEvalSubType::COMBINED, OperationMultipleEvalSubType::OCL,
  OperationMultipleEvalSubType::OCLFASTMP, OperationMultipleEvalSubType::OCLMP, OperationMultipleEvalSubType::OCLMASKMP, OperationMultipleEvalSubType::OCLOPT,
  OperationMultipleEvalSubType::OCLUNIFIED, OperationMultipleEvalSubType::CUDA
}
 
enum  OperationMultipleEvalType {
  OperationMultipleEvalType::DEFAULT, OperationMultipleEvalType::STREAMING, OperationMultipleEvalType::SUBSPACELINEAR, OperationMultipleEvalType::ADAPTIVE,
  OperationMultipleEvalType::MORTONORDER
}
 
enum  RefinementResultsUpdateType { ADDED_GRID_POINTS_LIST, DELETED_GRID_POINTS_LIST, SYSTEM_MATRIX_DECOMPOSITION }
 The type of message received in a UPDATE_GRID message type. More...
 
enum  RegularizationType {
  RegularizationType::Identity, RegularizationType::Laplace, RegularizationType::Diagonal, RegularizationType::Lasso,
  RegularizationType::ElasticNet, RegularizationType::GroupLasso
}
 
enum  ScorerMetricType { ScorerMetricType::mse }
 Enumeration of all supported metrics used to quantify approximation quality of a trained model. More...
 
enum  ScorerShufflingType { ScorerShufflingType::random, ScorerShufflingType::sequential }
 Enumeration of all supported shuffling types used to permute samples in a dataset. More...
 
enum  TaskType { TRAIN_FROM_BATCH, RECOMPUTE_SYSTEM_MATRIX_DECOMPOSITION }
 Used for the MPI Task Scheduler to differentiate between assigning tasks of different types. More...
 

Functions

DensityOCLMultiPlatform::OperationDensitycreateDensityOCLMultiPlatformConfigured (base::Grid &grid, size_t dimension, double lambda, base::OCLOperationConfiguration *parameters, size_t platform_id, size_t device_id)
 
DensityOCLMultiPlatform::OperationDensitycreateDensityOCLMultiPlatformConfigured (base::Grid &grid, size_t dimension, double lambda, std::string opencl_conf, size_t platform_id, size_t device_id)
 Generates opencl density multiplication operation given opencl device and configuration file. More...
 
DensityOCLMultiPlatform::OperationDensitycreateDensityOCLMultiPlatformConfigured (int *gridpoints, size_t gridsize, size_t dimension, double lambda, std::string opencl_conf, size_t platform_id, size_t device_id)
 Generates opencl density multiplication operation given opencl device and a serialized grid. More...
 
DensityOCLMultiPlatform::OperationDensitycreateDensityOCLMultiPlatformConfigured (int *gridpoints, size_t gridsize, size_t dimension, double lambda, base::OCLOperationConfiguration *parameters, size_t platform_id, size_t device_id)
 
DensityOCLMultiPlatform::OperationDensitycreateDensityOCLMultiPlatformConfigured (base::Grid &grid, size_t dimension, double lambda, std::string opencl_conf)
 Generates opencl density multiplication operation. More...
 
DensityOCLMultiPlatform::OperationCreateGraphOCLcreateNearestNeighborGraphConfigured (base::DataMatrix &dataset, size_t k, size_t dimensions, std::string opencl_conf, size_t platformid, size_t devicdeid)
 Generates the k nearest neighbors graph creation using a specific opencl device and a datamatrix. More...
 
DensityOCLMultiPlatform::OperationCreateGraphOCLcreateNearestNeighborGraphConfigured (double *dataset, size_t dataset_size, size_t k, size_t dimensions, sgpp::base::OCLOperationConfiguration *parameters, size_t platformid, size_t deviceid)
 
DensityOCLMultiPlatform::OperationCreateGraphOCLcreateNearestNeighborGraphConfigured (double *dataset, size_t dataset_size, size_t k, size_t dimensions, std::string opencl_conf, size_t platformid, size_t devicdeid)
 Generates the k nearest neighbors graph creation using a specific opencl device and a double vector. More...
 
DensityOCLMultiPlatform::OperationCreateGraphOCLcreateNearestNeighborGraphConfigured (base::DataMatrix &dataset, size_t k, size_t dimensions, std::string opencl_conf)
 Generates the k nearest neighbors graph creation. More...
 
base::OperationMultipleEvalcreateStreamingBSplineOCLConfigured (base::Grid &grid, base::DataMatrix &dataset, sgpp::datadriven::OperationMultipleEvalConfiguration &configuration)
 
base::OperationMultipleEvalcreateStreamingModOCLFastMultiPlatformConfigured (base::Grid &grid, base::DataMatrix &dataset, sgpp::datadriven::OperationMultipleEvalConfiguration &configuration)
 
base::OperationMultipleEvalcreateStreamingModOCLMaskMultiPlatformConfigured (base::Grid &grid, base::DataMatrix &dataset, sgpp::datadriven::OperationMultipleEvalConfiguration &configuration)
 
base::OperationMultipleEvalcreateStreamingModOCLOptConfigured (base::Grid &grid, base::DataMatrix &dataset, sgpp::datadriven::OperationMultipleEvalConfiguration &configuration)
 
base::OperationMultipleEvalcreateStreamingModOCLUnifiedConfigured (base::Grid &grid, base::DataMatrix &dataset, sgpp::datadriven::OperationMultipleEvalConfiguration &configuration)
 
base::OperationMultipleEvalcreateStreamingOCLMultiPlatformConfigured (base::Grid &grid, base::DataMatrix &dataset, sgpp::datadriven::OperationMultipleEvalConfiguration &configuration)
 Factory method for creating a new instance of this variant of OperationMultipleEval. More...
 
DensityOCLMultiPlatform::OperationPruneGraphOCLpruneNearestNeighborGraphConfigured (base::Grid &grid, size_t dimensions, base::DataVector &alpha, base::DataMatrix &data, double treshold, size_t k, std::string opencl_conf, size_t platformid, size_t deviceid)
 Generates the graph pruning operation for a specific opencl device. More...
 
DensityOCLMultiPlatform::OperationPruneGraphOCLpruneNearestNeighborGraphConfigured (int *gridpoints, size_t gridsize, size_t dimensions, double *alpha, base::DataMatrix &data, double treshold, size_t k, std::string opencl_conf, size_t platformid, size_t deviceid)
 Generates the graph pruning operation for a specific opencl device using a serialized grid. More...
 
DensityOCLMultiPlatform::OperationPruneGraphOCLpruneNearestNeighborGraphConfigured (int *gridpoints, size_t gridsize, size_t dimensions, double *alpha, base::DataMatrix &data, double treshold, size_t k, sgpp::base::OCLOperationConfiguration *parameters, size_t platformid, size_t deviceid)
 
DensityOCLMultiPlatform::OperationPruneGraphOCLpruneNearestNeighborGraphConfigured (base::Grid &grid, size_t dimensions, base::DataVector &alpha, base::DataMatrix &data, double treshold, size_t k, std::string opencl_conf)
 Generates the graph pruning operation. More...
 
template<class BASIS >
void test_calculateROCcurve (base::GridStorage *storage, BASIS &basis, base::DataVector &alpha, base::DataMatrix &data, base::DataVector &classes, base::DataVector &thresholds, base::DataMatrix &ROC_curve)
 Returns the number of correctly classified instances in data without boundaries. More...
 
template<class BASIS >
double test_dataset (base::GridStorage *storage, BASIS &basis, base::DataVector &alpha, base::DataMatrix &data, base::DataVector &classes)
 Returns the number of correctly classified instances in data without boundaries. More...
 
template<class BASIS >
double test_dataset_mse (base::GridStorage *storage, BASIS &basis, base::DataVector &alpha, base::DataMatrix &data, base::DataVector &refValues)
 Returns the MSE for given functions values at the evaluation points. More...
 
template<class BASIS >
double test_datasetWithCharacteristicNumber (base::GridStorage *storage, BASIS &basis, base::DataVector &alpha, base::DataMatrix &data, base::DataVector &classes, base::DataVector &charaNumbers, double threshold)
 Returns the number of correctly classified instances in data without boundaries. More...
 
void validate (boost::any &v, const std::vector< std::string > &values, sgpp::datadriven::OperationMultipleEvalType *target_type, int)
 
void validate (boost::any &v, const std::vector< std::string > &values, sgpp::datadriven::OperationMultipleEvalSubType *target_type, int)
 
void validate (boost::any &v, const std::vector< std::string > &values, sgpp::datadriven::LearnerMode *target_type, int)
 

Variables

double execTime_
 execution time More...
 
double GByte_
 number of transferred Gbytes More...
 
double GFlop_
 number of executed Floating Point operations More...
 
bool isRegression_
 is regression selected More...
 
bool isTrained_
 is the grid trained More...
 
bool isVerbose_
 

Typedef Documentation

typedef std::vector<std::pair<std::unique_ptr<DBMatOnlineDE>, double> > sgpp::datadriven::ClassDensityConntainer

Type definition to avoid typographical error in original learner.

Vector that holds level index pairs for every dimensions.

Enumeration Type Documentation

Supported file types for sgpp::datadriven::FileSampleProvider.

Enumerator
NONE 
ARFF 
CSV 
Enumerator
CG 
Decomposition 

Different fitter scenarios have different default values and support different operations.

Enumerator
RegressionLeastSquares 
DensityEstimation 
Enumerator
Float 
Double 
Enumerator
LEARN 
LEARNCOMPARE 
LEARNTEST 
Enumerator
LU 
Eigen 
Chol 
DenseIchol 
OrthoAdapt 

Different commands sent over MPI to allow the receiver to identify the message's contents.

Enumerator
NULL_COMMAND 

Used to identify packets where the command id has not been set.

UPDATE_GRID 

A packet that contains changes to the grid or the system matrix.

MERGE_GRID 

A packet that contains results from a training.

ASSIGN_BATCH 

A packet that assigns a batch with specified parameters to a worker.

COMPUTE_UPDATE_SYSTEM_MATRIX_DECOMPOSITION 

A packet that assigns updating the system matrix decomposition to a worker.

SHUTDOWN 

A packet that instructs a worker to shutdown its MPI facilities after completing all requests.

WORKER_SHUTDOWN_SUCCESS 

A confirmation packet sent by a worker to acknowledge all requests were completed.

Enumerator
NONE 
MASTERSLAVE 
HPX 
Enumerator
DEFAULT 
SIMPLE 
COMBINED 
OCL 
OCLFASTMP 
OCLMP 
OCLMASKMP 
OCLOPT 
OCLUNIFIED 
CUDA 
Enumerator
DEFAULT 
STREAMING 
SUBSPACELINEAR 
ADAPTIVE 
MORTONORDER 

The type of message received in a UPDATE_GRID message type.

Enumerator
ADDED_GRID_POINTS_LIST 

Packet contains a set of coordinates for newly created grid points.

DELETED_GRID_POINTS_LIST 

Packet contains a set of indices for grid points which were deleted.

SYSTEM_MATRIX_DECOMPOSITION 

Packet contains a part of an updated system matrix decomposition.

Enumerator
Identity 
Laplace 
Diagonal 
Lasso 
ElasticNet 
GroupLasso 

Enumeration of all supported metrics used to quantify approximation quality of a trained model.

An entry exists for each object that derives from sgpp::datadriven::Metric. Used for configuration and factory methods.

Enumerator
mse 

Enumeration of all supported shuffling types used to permute samples in a dataset.

An entry exists for each object that derives from sgpp::datadriven::ShufflingFunctor. Used for configuration and factory methods.

Enumerator
random 
sequential 

Used for the MPI Task Scheduler to differentiate between assigning tasks of different types.

Enumerator
TRAIN_FROM_BATCH 
RECOMPUTE_SYSTEM_MATRIX_DECOMPOSITION 

Function Documentation

sgpp::datadriven::DensityOCLMultiPlatform::OperationDensity * sgpp::datadriven::createDensityOCLMultiPlatformConfigured ( base::Grid grid,
size_t  dimension,
double  lambda,
std::string  opencl_conf,
size_t  platform_id,
size_t  device_id 
)

Generates opencl density multiplication operation given opencl device and configuration file.

References grid(), lambda, and sgpp::datadriven::DensityOCLMultiPlatform::OperationDensity::load_default_parameters().

sgpp::datadriven::DensityOCLMultiPlatform::OperationDensity * sgpp::datadriven::createDensityOCLMultiPlatformConfigured ( int *  gridpoints,
size_t  gridsize,
size_t  dimension,
double  lambda,
std::string  opencl_conf,
size_t  platform_id,
size_t  device_id 
)

Generates opencl density multiplication operation given opencl device and a serialized grid.

References lambda, and sgpp::datadriven::DensityOCLMultiPlatform::OperationDensity::load_default_parameters().

DensityOCLMultiPlatform::OperationDensity * sgpp::datadriven::createDensityOCLMultiPlatformConfigured ( int *  gridpoints,
size_t  gridsize,
size_t  dimension,
double  lambda,
base::OCLOperationConfiguration parameters,
size_t  platform_id,
size_t  device_id 
)
sgpp::datadriven::DensityOCLMultiPlatform::OperationDensity * sgpp::datadriven::createDensityOCLMultiPlatformConfigured ( base::Grid grid,
size_t  dimension,
double  lambda,
std::string  opencl_conf 
)
DensityOCLMultiPlatform::OperationCreateGraphOCL * sgpp::datadriven::createNearestNeighborGraphConfigured ( base::DataMatrix dataset,
size_t  k,
size_t  dimensions,
std::string  opencl_conf,
size_t  platformid,
size_t  deviceid 
)
DensityOCLMultiPlatform::OperationCreateGraphOCL * sgpp::datadriven::createNearestNeighborGraphConfigured ( double *  dataset,
size_t  dataset_size,
size_t  k,
size_t  dimensions,
sgpp::base::OCLOperationConfiguration parameters,
size_t  platformid,
size_t  deviceid 
)
DensityOCLMultiPlatform::OperationCreateGraphOCL * sgpp::datadriven::createNearestNeighborGraphConfigured ( double *  dataset,
size_t  dataset_size,
size_t  k,
size_t  dimensions,
std::string  opencl_conf,
size_t  platformid,
size_t  deviceid 
)

Generates the k nearest neighbors graph creation using a specific opencl device and a double vector.

References dataset, and sgpp::datadriven::DensityOCLMultiPlatform::OperationCreateGraphOCL::load_default_parameters().

DensityOCLMultiPlatform::OperationCreateGraphOCL * sgpp::datadriven::createNearestNeighborGraphConfigured ( base::DataMatrix dataset,
size_t  k,
size_t  dimensions,
std::string  opencl_conf 
)
base::OperationMultipleEval * sgpp::datadriven::createStreamingOCLMultiPlatformConfigured ( base::Grid grid,
base::DataMatrix dataset,
sgpp::datadriven::OperationMultipleEvalConfiguration configuration 
)

Factory method for creating a new instance of this variant of OperationMultipleEval.

This factory method configures the resulting object by using the parameters of the provided configuration. If no parameters are provided, the default parameter values are used. If a configuration is provided, but some entries were not set, this class adds the missing entries with their default values. This class is the non-templated entry point for the templated inner objects. Templates are used to implement different floating point precision.

See also
OperationMultiEvalStreamingOCLMultiPlatform
Parameters
gridThe sparse grid to evaluate
datasetThe datapoints to evaluate
configurationConfiguration that may contain a parameter object for configuration details

References sgpp::datadriven::StreamingOCLMultiPlatform::Configuration::augmentDefaultParameters(), dataset, sgpp::datadriven::OperationMultipleEvalConfiguration::getParameters(), and grid().

Referenced by sgpp::op_factory::createOperationMultipleEval().

sgpp::datadriven::DensityOCLMultiPlatform::OperationPruneGraphOCL * sgpp::datadriven::pruneNearestNeighborGraphConfigured ( base::Grid grid,
size_t  dimensions,
base::DataVector alpha,
base::DataMatrix data,
double  treshold,
size_t  k,
std::string  opencl_conf,
size_t  platformid,
size_t  deviceid 
)
sgpp::datadriven::DensityOCLMultiPlatform::OperationPruneGraphOCL * sgpp::datadriven::pruneNearestNeighborGraphConfigured ( int *  gridpoints,
size_t  gridsize,
size_t  dimensions,
double *  alpha,
base::DataMatrix data,
double  treshold,
size_t  k,
std::string  opencl_conf,
size_t  platformid,
size_t  deviceid 
)

Generates the graph pruning operation for a specific opencl device using a serialized grid.

References alpha, and sgpp::datadriven::DensityOCLMultiPlatform::OperationPruneGraphOCL::load_default_parameters().

DensityOCLMultiPlatform::OperationPruneGraphOCL * sgpp::datadriven::pruneNearestNeighborGraphConfigured ( int *  gridpoints,
size_t  gridsize,
size_t  dimensions,
double *  alpha,
base::DataMatrix data,
double  treshold,
size_t  k,
sgpp::base::OCLOperationConfiguration parameters,
size_t  platformid,
size_t  deviceid 
)
sgpp::datadriven::DensityOCLMultiPlatform::OperationPruneGraphOCL * sgpp::datadriven::pruneNearestNeighborGraphConfigured ( base::Grid grid,
size_t  dimensions,
base::DataVector alpha,
base::DataMatrix data,
double  treshold,
size_t  k,
std::string  opencl_conf 
)
template<class BASIS >
void sgpp::datadriven::test_calculateROCcurve ( base::GridStorage storage,
BASIS &  basis,
base::DataVector alpha,
base::DataMatrix data,
base::DataVector classes,
base::DataVector thresholds,
base::DataMatrix ROC_curve 
)
template<class BASIS >
double sgpp::datadriven::test_dataset ( base::GridStorage storage,
BASIS &  basis,
base::DataVector alpha,
base::DataMatrix data,
base::DataVector classes 
)
template<class BASIS >
double sgpp::datadriven::test_datasetWithCharacteristicNumber ( base::GridStorage storage,
BASIS &  basis,
base::DataVector alpha,
base::DataMatrix data,
base::DataVector classes,
base::DataVector charaNumbers,
double  threshold 
)
void sgpp::datadriven::validate ( boost::any &  v,
const std::vector< std::string > &  values,
sgpp::datadriven::OperationMultipleEvalType target_type,
int   
)

References DEFAULT, STREAMING, and SUBSPACELINEAR.

void sgpp::datadriven::validate ( boost::any &  v,
const std::vector< std::string > &  values,
sgpp::datadriven::OperationMultipleEvalSubType target_type,
int   
)
void sgpp::datadriven::validate ( boost::any &  v,
const std::vector< std::string > &  values,
sgpp::datadriven::LearnerMode target_type,
int   
)

References LEARN, LEARNCOMPARE, and LEARNTEST.

Variable Documentation

double sgpp::datadriven::execTime_

execution time

double sgpp::datadriven::GByte_

number of transferred Gbytes

double sgpp::datadriven::GFlop_

number of executed Floating Point operations

bool sgpp::datadriven::isRegression_

is regression selected

bool sgpp::datadriven::isTrained_

is the grid trained

bool sgpp::datadriven::isVerbose_