Java* API Reference for Intel® Data Analytics Acceleration Library 2019 Update 4
Provides methods to train models that depend on the data provided in the distributed processing mode. For example, these methods enable training the linear regression model. Classes that implement specific algorithms of model training in the distributed processing mode are derived classes of the TrainingDistributed class. The class additionally provides methods for validation of input and output parameters of the algorithms. More...
TrainingDistributed | ( | DaalContext | context | ) |
Constructs the training algorithm in the distributed processing mode
context | Context to manage the training algorithm in the distributed processing mode |
void checkComputeParams | ( | ) |
Validates parameters of the compute method
void checkFinalizeComputeParams | ( | ) |
Validates parameters of the finalizeCompute method
|
abstract |
Returns the newly allocated training algorithm with a copy of input objects and parameters of this algorithm
context | Context to manage the training algorithm |
PartialResult compute | ( | ) |
Computes partial results of the algorithm in the distributed processing mode
void dispose | ( | ) |
Releases memory allocated for the native algorithm object
Implements Disposable.
Result finalizeCompute | ( | ) |
Computes final results of the algorithm using partial results in the distributed processing mode.
For more complete information about compiler optimizations, see our Optimization Notice.