Java* API Reference for Intel® Data Analytics Acceleration Library 2019

List of all members

Parameters of the LBFGS algorithm. More...

Class Constructor

Parameter ( DaalContext  context)

Constructs the parameters of the LBFGS algorithm

Parameters
contextContext to manage the parameters of the LBFGS algorithm
Parameter ( DaalContext  context,
long  cObject 
)

Constructs the parameter for the LBFGS algorithm

Parameters
contextContext to manage the LBFGS algorithm
cObjectPointer to C++ implementation of the parameter

Detailed Description

Member Function Documentation

NumericTable getBatchIndices ( )

Gets the numeric table that represents 32 bit integer indices of terms of the objective function.

Returns
The numeric table that represents 32 bit integer indices of terms of the objective function
NumericTable getCorrectionPairBatchIndices ( )

Gets the numeric table that represents 32 bit integer indices of terms of the objective function.

Returns
The numeric table that represents 32 bit integer indices of terms of the objective function
long getCorrectionPairBatchSize ( )

Gets the number of observations to compute the sub-sampled Hessian for correction pairs computation

Returns
The number of observations to compute the sub-sampled Hessian for correction pairs computation
long getL ( )

Gets the number of iterations between the curvature estimates calculations

Returns
The number of iterations between the curvature estimates calculations
long getM ( )

Gets the memory parameter of the LBFGS algorithm. Which is the maximum number of correction pairs that define the approximation of inverse Hessian matrix

Returns
The memory parameter of the LBFGS algorithm
long getSeed ( )
Deprecated:
This item will be removed in a future release.

Gets the seed for random generation of 32 bit integer indices of terms in the objective function.

Returns
The seed for random generation of 32 bit integer indices of terms in the objective function.
NumericTable getStepLengthSequence ( )

Gets the numeric table that contains values of the step-length sequence

Returns
The numeric table that contains values of the step-length sequence
void setBatchIndices ( NumericTable  batchIndices)

Sets the numeric table of size nIterations x batchSize that represent indices that will be used instead of random values for the stochastic gradient computations. If no indices are provided, the implementation will generate random indices.

Parameters
batchIndicesThe numeric table that represents 32 bit integer indices of terms of the objective function
void setCorrectionPairBatchIndices ( NumericTable  batchIndices)

Sets the numeric table of size (nIterations / L) x correctionPairBatchSize that represent indices that will be used instead of random values for the sub-sampled Hessian matrix computations. If no indices are provided, the implementation will generate random indices.

Parameters
batchIndicesThe numeric table that represents 32 bit integer indices of terms of the objective function
void setCorrectionPairBatchSize ( long  batchSize)

Sets the number of observations to compute the sub-sampled Hessian for correction pairs computation

Parameters
batchSizeThe number of observations to compute the sub-sampled Hessian for correction pairs computation
void setEngine ( com.intel.daal.algorithms.engines.BatchBase  engine)

Sets the engine to be used by the algorithm

Parameters
engineto be used by the algorithm
void setL ( long  L)

Sets the number of iterations between the curvature estimates calculations

Parameters
LThe number of iterations between the curvature estimates calculations
void setM ( long  m)

Sets the memory parameter of the LBFGS algorithm. Which is the maximum number of correction pairs that define the approximation of inverse Hessian matrix

Parameters
mThe memory parameter of the LBFGS algorithm
void setSeed ( long  seed)
Deprecated:
This item will be removed in a future release.

Sets the seed for random generation of 32 bit integer indices of terms in the objective function.

Parameters
seedThe seed for random generation of 32 bit integer indices of terms in the objective function.
void setStepLengthSequence ( NumericTable  stepLengthSequence)

Sets the numeric table that contains values of the step-length sequence

Parameters
stepLengthSequenceThe numeric table that contains values of the step-length sequence

The documentation for this class was generated from the following file:

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