DecisionTreeTrain Method (Matrix, IntegerArray, IListDecisionTreeFactorType, Vector, Double, Int32, Boolean, Boolean, Boolean, RandomGenerator)

public void Train(
	Matrix observations,
	IntegerArray classes,
	IList<DecisionTreeFactorType> factorTypes = null,
	Vector weights = null,
	double minimumObservationsNumber = 2,
	int numFoldsPruning = 5,
	bool doPrune = true,
	bool useHeuristics = true,
	bool useOneSERule = false,
	RandomGenerator random = null
)

Parameters

observations

Type: FinMath.LinearAlgebraMatrix
Observation matrix (N*M size). Columns corresponds to factors, rows to observation.

classes

Type: FinMath.DataStructuresIntegerArray
The vector (N size) class types.

factorTypes (Optional)

Type: System.Collections.GenericIListDecisionTreeFactorType
The array (M size) factor types (Continuous by default).

weights (Optional)

Type: FinMath.LinearAlgebraVector
The vector (N size) of any observation weight (ones by default).

minimumObservationsNumber (Optional)

Type: SystemDouble
Minimum number of observations in at the terminal nodes (2 by default).

numFoldsPruning (Optional)

Type: SystemInt32
Number of folds for minimal cost-complexity pruning (5 by default).

doPrune (Optional)

Type: SystemBoolean
Is use minimal cost-complexity pruning (default true).

useHeuristics (Optional)

Type: SystemBoolean
Is use heuristic search for nominal factors in multi-class problems (default true).

useOneSERule (Optional)

Type: SystemBoolean
If use the 1SE rule to make final decision tree - prune thee more harsh (default false).

random (Optional)

Type: FinMath.StatisticsRandomGenerator
Random number generator for cross-validation.

Reference