Class for building model of the gradient boosted trees classification algorithm
More...
Constructs the gradient boosted trees classification model builder
- Parameters
-
context | Context to manage gradient boosted trees classification model builder |
nFeatures | Number of features in training data |
nIterations | Number of trees in model for each class |
nClasses | Number of classes |
long addLeafNode |
( |
long |
treeId, |
|
|
long |
parentId, |
|
|
long |
position, |
|
|
double |
response |
|
) |
| |
Create Leaf node and add it to certain tree
- Parameters
-
treeId | Tree to which new node is added |
parentId | Parent node to which new node is added (use noParent for root node) |
position | Position in parent (e.g. 0 for left and 1 for right child in a binary tree) |
response | Response 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
-
treeId | Tree to which new node is added |
parentId | Parent node to which new node is added (use noParent for root node) |
position | Position in parent (e.g. 0 for left and 1 for right child in a binary tree) |
featureIndex | Feature index for spliting |
featureValue | Feature value for spliting |
- Returns
- Positive number node identifier
long createTree |
( |
long |
nNodes, |
|
|
long |
classLabel |
|
) |
| |
Create certain tree in the gradient boosted trees classification model for certain class
- Parameters
-
nNodes | Number of nodes in created tree |
classLabel | Label of class for which tree is created. classLabel bellows interval from 0 to (nClasses - 1) |
- Returns
- Positive number tree identifier
Get built model of gradient boosted trees classification
- Returns
- Model of gradient boosted trees classification
The documentation for this class was generated from the following file:
- gbt/classification/ModelBuilder.java