DecisionNode struct
#include <DecisionNode.h>

DecisionNodes can either be branches or leafs. Branch nodes store data used to instruct the traversal of the decision tree. The data used for traversal are a column index of a DataFrame and a row value. Both of these are used for comparison with new data to go left (true) or right (false). Leaf nodes contain data used for prediction.

Contents

Constructors, destructors, conversion operators

DecisionNode(double splitLoss, std::vector<double> values, int splitColumn = 0, Generic* splitValue = nullptr, DecisionNode* left = nullptr, DecisionNode* right = nullptr)
Construct a new DecisionNode object.

Public functions

auto isLeaf() -> bool
Determines whethers node is a leaf node.

Public variables

DecisionNode* left
DecisionNode* right
int splitColumn
Generic* splitValue
double splitLoss
int sampleSize
std::vector<double> values

Function documentation

DecisionNode::DecisionNode(double splitLoss, std::vector<double> values, int splitColumn = 0, Generic* splitValue = nullptr, DecisionNode* left = nullptr, DecisionNode* right = nullptr)

Construct a new DecisionNode object.

Parameters
splitLoss Used for optimizing the fit of a decision tree.
values An array containing the expected value (mean) or the counts of each class for DecisionTreeRegressor and DecisionTreeClassifier, respectively.
splitColumn The column index used for splitting the node.
splitValue The row value used for splitting the node.
left The node that represents true for comparison.
right The node that represents false for comparison.

bool DecisionNode::isLeaf()

Determines whethers node is a leaf node.

Returns true a leaf node.