Mojo struct

DecisionTree

@memory_only
struct DecisionTree[criterion: String = "gini"]

A decision tree supporting both classification and regression.

Aliases

• loss_func = gini if (criterion == String("gini")) else mse_loss if (criterion == String("mse")) else entropy
• c_precompute = gini_precompute if (criterion == String("gini")) else entropy_precompute
• r_precompute: __mlir_type.!kgen.func.literal<:!lit.fn<[2]("size": !lit.struct<_std::_builtin::_simd::_SIMD<:!lit.struct<_std::_builtin::_dtype::_DType> {:dtype f32}, :!lit.struct<_std::_builtin::_int::_Int> {1}>>, "sum": !lit.struct<_std::_builtin::_simd::_SIMD<:!lit.struct<_std::_builtin::_dtype::_DType> {:dtype f32}, :!lit.struct<_std::_builtin::_int::_Int> {1}>>, "sum_sq": !lit.struct<_std::_builtin::_simd::_SIMD<:!lit.struct<_std::_builtin::_dtype::_DType> {:dtype f32}, :!lit.struct<_std::_builtin::_int::_Int> {1}>>, ?, "error": !lit.ref<!lit.struct<_std::_builtin::_error::_Error>, mut [0,0]> byref_error, "result": !lit.ref<:meta<!lit.struct<_std::_builtin::_simd::_SIMD<:!lit.struct<_std::_builtin::_dtype::_DType> {:dtype f32}, :!lit.struct<_std::_builtin::_int::_Int> {1}>>> sugar_alias("Float320x21", _std::_builtin::_simd::_SIMD<:!lit.struct<_std::_builtin::_dtype::_DType> {:dtype f32}, :!lit.struct<_std::_builtin::_int::_Int> {1}>), mut *[0,1]> byref_result) throws -> i1> #kgen.func.symbol<_mojmelo::_utils::_utils::_"mse_loss_precompute(::SIMD[::DType(float32), ::Int(1)],::SIMD[::DType(float32), ::Int(1)],::SIMD[::DType(float32), ::Int(1)])">> = fn_literal`
• MODEL_ID = 9

Parameters

• criterion (String): The function to measure the quality of a split: For classification -> 'entropy', 'gini'; For regression -> 'mse'.

Fields

• min_samples_split (Int): The minimum number of samples required to split an internal node.
• max_depth (Int): The maximum depth of the tree.
• n_feats (Int): The number of features to consider when looking for the best split.
• root (Optional[UnsafePointer[Node, MutAnyOrigin]])

Implemented traits

AnyType, CV, Copyable, ImplicitlyCopyable, ImplicitlyDestructible, Movable

Methods

__init__

fn __init__(out self, min_samples_split: Int = 2, max_depth: Int = 100, n_feats: Int = -1, random_state: Int = 42)

Args:

• min_samples_split (Int)
• max_depth (Int)
• n_feats (Int)
• random_state (Int)
• self (Self)

Returns:

Self

fn __init__(out self, params: Dict[String, String])

Args:

• params (Dict[String, String])
• self (Self)

Returns:

Self

Raises:

__del__

fn __del__(deinit self)

Args:

• self (Self)

fit

fn fit(mut self, X: Matrix, y: Matrix)

Build a decision tree from the training set.

Args:

• self (Self)
• X (Matrix)
• y (Matrix)

Raises:

fit_weighted

fn fit_weighted(mut self, X: Matrix, y_with_weights: Matrix)

Build a decision tree from a weighted training set.

Args:

• self (Self)
• X (Matrix)
• y_with_weights (Matrix)

Raises:

predict

fn predict(self, X: Matrix) -> Matrix

Predict class or regression value for X.

Args:

• self (Self)
• X (Matrix)

Returns:

Matrix: The predicted values.

Raises:

save

fn save(self, path: String)

Save model data necessary for prediction to the specified path.

Args:

• self (Self)
• path (String)

Raises:

load

@staticmethod
fn load(path: String) -> Self

Load a saved model from the specified path for prediction.

Args:

• path (String)

Returns:

Self

Raises: