Java* API Reference for Intel® Data Analytics Acceleration Library 2019 Update 5

List of all members
ModelBuilder Class Reference

Class for building model of the gradient boosted trees regression algorithm More...

Class Constructor

ModelBuilder ( DaalContext  context,
long  nFeatures,
long  nIterations 
)

Constructs the gradient boosted trees regression model builder

Parameters
contextContext to manage gradient boosted trees regression model builder
nFeaturesNumber of features in training data
nIterationsNumber of trees in model and iterations performed on training stage of gbt algorithm

Detailed Description

Member Function Documentation

long addLeafNode ( long  treeId,
long  parentId,
long  position,
double  response 
)

Create Leaf node and add it to certain tree

Parameters
treeIdTree to which new node is added
parentIdParent node to which new node is added (use noParent for root node)
positionPosition in parent (e.g. 0 for left and 1 for right child in a binary tree)
responseResponse value for leaf node to be predicted
Returns
Positive number node identifier
long addSplitNode ( long  treeId,
long  parentId,
long  position,
long  featureIndex,
double  featureValue 
)

Create Split node and add it to certain tree

Parameters
treeIdTree to which new node is added
parentIdParent node to which new node is added (use noParent for root node)
positionPosition in parent (e.g. 0 for left and 1 for right child in a binary tree)
featureIndexFeature index for spliting
featureValueFeature value for spliting
Returns
Positive number node identifier
long createTree ( long  nNodes)

Create certain tree in the gradient boosted trees regression model

Parameters
nNodesNumber of nodes in created tree
Returns
Positive number tree identifier
Model getModel ( )

Get built model of gradient boosted trees regression

Returns
Model of gradient boosted trees regression

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

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