C++ API Reference for Intel® Data Analytics Acceleration Library 2020 Update 1

References | Namespaces | Classes | Enumerations | Functions | Variables

Contains a class for lasso regression model-based training. More...

References

 Batch
 

Namespaces

 daal::algorithms::lasso_regression::training
 Contains a class for lasso regression model-based training.
 
 daal::algorithms::lasso_regression::training::interface1
 Contains version 1.0 of the Intel(R) Data Analytics Acceleration Library (Intel(R) DAAL) interface.
 

Classes

struct  Parameter
 Parameters for the lasso regression algorithm. More...
 
class  InputIface
 Abstract class that specifies the interface of input objects for lasso regression model-based training. More...
 
class  Input
 Input objects for lasso regression model-based training More...
 
class  Result
 Provides methods to access the result obtained with the compute() method of lasso regression model-based training. More...
 

Enumerations

enum  Method { defaultDense = 0 }
 Computation methods for lasso regression model-based training. More...
 
enum  InputId { data = linear_model::training::data, dependentVariables = linear_model::training::dependentVariables }
 Available identifiers of input objects for lasso regression model-based training. More...
 
enum  OptionalInputId { optionalArgument = lastInputId + 1 }
 
enum  OptionalDataId { weights, gramMatrix }
 
enum  ResultToComputeId { computeGramMatrix = 0x00000001ULL }
 
enum  ResultId { model = linear_model::training::model }
 Available identifiers of the result of lasso regression model-based training. More...
 
enum  OptionalResultNumericTableId { gramMatrixId = lastResultId + 1 }
 
enum  DataUseInComputation { doNotUse = 0, doUse = 1 }
 Available identifiers for input data corruption. More...
 

Functions

virtual size_t getNumberOfFeatures () const =0
 
virtual size_t getNumberOfDependentVariables () const =0
 
 Input ()
 
data_management::NumericTablePtr get (InputId id) const
 
void set (InputId id, const data_management::NumericTablePtr &value)
 
algorithms::OptionalArgumentPtr get (OptionalInputId id) const
 
void set (OptionalInputId id, const algorithms::OptionalArgumentPtr &ptr)
 
data_management::NumericTablePtr get (OptionalDataId id) const
 
void set (OptionalDataId id, const data_management::NumericTablePtr &ptr)
 
size_t getNumberOfFeatures () const DAAL_C11_OVERRIDE
 
size_t getNumberOfDependentVariables () const DAAL_C11_OVERRIDE
 
services::Status check (const daal::algorithms::Parameter *par, int method) const DAAL_C11_OVERRIDE
 
lasso_regression::ModelPtr get (ResultId id) const
 
void set (ResultId id, const lasso_regression::ModelPtr &value)
 
data_management::NumericTablePtr get (OptionalResultNumericTableId id) const
 
void set (OptionalResultNumericTableId id, const data_management::NumericTablePtr &value)
 
template<typename algorithmFPType >
DAAL_EXPORT services::Status allocate (const daal::algorithms::Input *input, const Parameter *parameter, const int method)
 
services::Status check (const daal::algorithms::Input *input, const daal::algorithms::Parameter *par, int method) const DAAL_C11_OVERRIDE
 

Variables

data_management::NumericTablePtr lassoParameters
 
SolverPtr optimizationSolver
 
DataUseInComputation dataUseInComputation
 
DAAL_UINT64 optResultToCompute
 

Enumeration Type Documentation

enum DataUseInComputation

Enumerator
doNotUse 

The input data and labels can`t be corrupted

doUse 

The input data and labels can be corrupted

enum InputId

Enumerator
data 

Input data table

dependentVariables 

Values of the dependent variable for the input data

enum Method

Enumerator
defaultDense 

Normal equations method

enum OptionalDataId

Available identifiers of optional input for the iterative solver

Enumerator
weights 

NumericTable of size 1 x n with weights of samples. Applied for all method

gramMatrix 

NumericTable of size p x p with last iteration number. Applied for all method

enum OptionalInputId

Available identifiers of optional input for the iterative solver

Enumerator
optionalArgument 

Algorithm-specific input data, can be generated by previous runs of the algorithm

enum OptionalResultNumericTableId

Available identifiers of results obtained in the training stage of the regression algorithm

Enumerator
gramMatrixId 

Numeric table of size: p x p, containing computed Gram matrix

enum ResultId

Enumerator
model 

Lasso regression model

enum ResultToComputeId

Available identifiers to specify the result to compute

Enumerator
computeGramMatrix 

The flag to compute Gram Matrix

Function Documentation

DAAL_EXPORT services::Status allocate ( const daal::algorithms::Input *  input,
const Parameter parameter,
const int  method 
)

Allocates memory to store the result of lasso regression model-based training

Parameters
[in]inputPointer to an object containing the input data
[in]parameterParameter of lasso regression model-based training
[in]methodComputation method for the algorithm
Returns
Status of computations
services::Status check ( const daal::algorithms::Parameter *  par,
int  method 
) const

Checks an input object for the lasso regression algorithm

Parameters
[in]parAlgorithm parameter
[in]methodComputation method
Returns
Status of computations
services::Status check ( const daal::algorithms::Input *  input,
const daal::algorithms::Parameter *  par,
int  method 
) const

Checks the result of lasso regression model-based training

Parameters
[in]inputInput object for the algorithm
[in]parParameter of the algorithm
[in]methodComputation method
Returns
Status of computations
data_management::NumericTablePtr get ( InputId  id) const

Returns an input object for lasso regression model-based training

Parameters
[in]idIdentifier of the input object
Returns
Input object that corresponds to the given identifier
algorithms::OptionalArgumentPtr get ( OptionalInputId  id) const

Returns optional input of the iterative solver algorithm

Parameters
[in]idIdentifier of the optional input data
Returns
Input data that corresponds to the given identifier
data_management::NumericTablePtr get ( OptionalDataId  id) const

Returns input NumericTable containing optional data

Parameters
[in]idIdentifier of the input numeric table
Returns
Input numeric table that corresponds to the given identifier
lasso_regression::ModelPtr get ( ResultId  id) const

Returns the result of lasso regression model-based training

Parameters
[in]idIdentifier of the result
Returns
Result that corresponds to the given identifier
data_management::NumericTablePtr get ( OptionalResultNumericTableId  id) const

Returns the result of lasso regression model-based training

Parameters
[in]idIdentifier of the result
Returns
Result that corresponds to the given identifier
virtual size_t getNumberOfDependentVariables ( ) const
pure virtual

Returns the number of dependent variables

Returns
Number of dependent variables

Implemented in Input.

size_t getNumberOfDependentVariables ( ) const
virtual

Returns the number of dependent variables

Returns
Number of dependent variables

Implements InputIface.

virtual size_t getNumberOfFeatures ( ) const
pure virtual

Returns the number of columns in the input data set

Returns
Number of columns in the input data set

Implemented in Input.

size_t getNumberOfFeatures ( ) const
virtual

Returns the number of columns in the input data set

Returns
Number of columns in the input data set

Implements InputIface.

Input ( )

Default constructor

void set ( InputId  id,
const data_management::NumericTablePtr &  value 
)

Sets an input object for lasso regression model-based training

Parameters
[in]idIdentifier of the input object
[in]valuePointer to the object
void set ( OptionalInputId  id,
const algorithms::OptionalArgumentPtr &  ptr 
)

Sets optional input for the iterative solver algorithm

Parameters
[in]idIdentifier of the input object
[in]ptrPointer to the object
void set ( OptionalDataId  id,
const data_management::NumericTablePtr &  ptr 
)

Sets optional input for the algorithm

Parameters
[in]idIdentifier of the input object
[in]ptrPointer to the object
void set ( ResultId  id,
const lasso_regression::ModelPtr &  value 
)

Sets the result of lasso regression model-based training

Parameters
[in]idIdentifier of the result
[in]valueResult
void set ( OptionalResultNumericTableId  id,
const data_management::NumericTablePtr &  value 
)

Sets the result of lasso regression model-based training

Parameters
[in]idIdentifier of the input object
[in]valueInput object

Variable Documentation

DataUseInComputation dataUseInComputation

The flag allows to corrupt input data

data_management::NumericTablePtr lassoParameters

Numeric table that contains values of lasso parameters

SolverPtr optimizationSolver

Default is coordinate descent solver

DAAL_UINT64 optResultToCompute

64 bit integer flag that indicates the optional results to compute

For more complete information about compiler optimizations, see our Optimization Notice.