结构体 simple_ml.base.base_model

二叉树节点

class BinaryTreeNode:

    def __init__(self, left=None, right=None, data_id=None, 
    feature_id=None, value=None, leaf_label=None):
        if data_id is None:
            data_id = []
        self.left = left
        self.right = right
        self.data_id = data_id
        self.feature_id = feature_id
        self.value = value
        self.leaf_label = leaf_label
  名称 类型 描述
Parameters: left BinaryTreeNode 左节点
  right BinaryTreeNode 右节点
  data_id List or np.array 样本id列表
  feature_id int 特征id
  value float or int 对应特征的值
  leaf_label float or int 叶子节点的标签（非叶子节点为None)
Returns:   Void  

多叉树节

class MultiTreeNode:

    def __init__(self, child=None, data_id=None, feature_id=None, value=None, leaf_label=None):
        if data_id is None:
            data_id = []
        self.child = child
        self.data_id = data_id
        self.feature_id = feature_id
        self.value = value
        self.leaf_label = leaf_label
  名称 类型 描述
Parameters: child List[MultiTreeNode] 孩子节点列表
  data_id List or np.array 样本id列表
  feature_id int 特征id
  value float or int 对应特征的值
  leaf_label float or int 叶子节点的标签（非叶子节点为None)
Returns:   Void  

GBDT二叉树节点

from simple_ml.base.base_model import BinaryTreeNode

class GBDTTreeNode(BinaryTreeNode):

    def __init__(self, left=None, right=None, data_id=None, feature_id=None, value=None, leaf_label=None):
        super(GBDTTreeNode, self).__init__(left, right, data_id, feature_id, value, leaf_label)
        self.gamma = None

继承BinaryTreeNode

gamma值用以记录每个叶子节点的最优值

  名称 类型 描述
Parameters: left BinaryTreeNode 左节点
  right BinaryTreeNode 右节点
  data_id List or np.array 样本id列表
  feature_id int 特征id
  value float or int 对应特征的值
  leaf_label float or int 叶子节点的标签（非叶子节点为None)
  gamma float or int 叶子节点的最优值
Returns:   Void  

返回主页