Struct RandomForestClassifier 

pub struct RandomForestClassifier { /* private fields */ }

Random-forest classifier using bootstrap aggregation of CART classifiers.

The default is 100 bootstrap trees with max_features = sqrt(n_features) at each split, matching scikit-learn’s RandomForestClassifier defaults. random_state defaults to zero so fits are reproducible; call with_random_state to choose another seed.

use datarust::ensemble::RandomForestClassifier;
use datarust::traits::{PredictProba, Predictor};
use datarust::Matrix;

let x = Matrix::new(vec![
    vec![0.0, 0.0], vec![0.0, 1.0],
    vec![1.0, 0.0], vec![1.0, 1.0],
])?;
let y = vec![0.0, 1.0, 1.0, 0.0];
let mut forest = RandomForestClassifier::new()
    .with_n_estimators(25)
    .with_random_state(7);
forest.fit(&x, &y)?;
let predictions = forest.predict(&x)?;
let probabilities = forest.predict_proba(&x)?;
assert_eq!(predictions.len(), 4);
assert_eq!(probabilities.ncols(), 2);

Implementations

impl RandomForestClassifier

pub fn new() -> Self

Creates a classifier with scikit-learn-style defaults.

pub fn with_n_estimators(self, n: usize) -> Self

Sets the number of trees.

pub fn with_criterion(self, criterion: Criterion) -> Self

Sets the tree split criterion (Gini or Entropy). Criterion::SquaredError is regression-only and is rejected by fit.

pub fn with_max_depth(self, depth: Option<usize>) -> Self

Sets the maximum depth of each tree (None means unlimited).

pub fn with_min_samples_split(self, n: usize) -> Self

Sets the minimum samples required to split a node.

pub fn with_min_samples_leaf(self, n: usize) -> Self

Sets the minimum samples required in each leaf.

pub fn with_min_impurity_decrease(self, value: f64) -> Self

Sets the minimum impurity decrease required for a split.

pub fn with_max_features(self, value: MaxFeatures) -> Self

Sets the number of features considered at each split.

pub fn with_bootstrap(self, bootstrap: bool) -> Self

Enables or disables bootstrap sampling.

pub fn with_random_state(self, seed: u64) -> Self

Sets the deterministic forest seed.

pub fn n_estimators(&self) -> usize

Number of fitted trees.

pub fn n_features_in(&self) -> usize

Number of features seen during fitting.

pub fn trees(&self) -> &[DecisionTreeClassifier]

Borrow the fitted decision trees.

pub fn classes(&self) -> Option<&[f64]>

Sorted original class labels.

pub fn estimators_samples(&self) -> Vec<Vec<usize>>

Regenerates the row indices used to fit each tree (scikit-learn’s estimators_samples_ semantics). With bootstrap enabled, each inner vector contains n_samples draws with replacement; with bootstrap disabled it is the original row order. Indices are derived from the stored per-tree seeds on demand, so the fitted model does not retain n_estimators × n_samples indices.

pub fn feature_importances(&self) -> Result<Vec<f64>>

Mean normalized impurity reduction across the non-trivial fitted trees, matching sklearn’s feature_importances_ behavior.

Trait Implementations

impl Classifier for RandomForestClassifier

fn score(&self, x: &Matrix, y: &[f64]) -> Result<f64>

Mean classification accuracy, mirroring sklearn’s default classifier score.

impl Clone for RandomForestClassifier

fn clone(&self) -> RandomForestClassifier

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for RandomForestClassifier

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for RandomForestClassifier

fn default() -> Self

Returns the “default value” for a type.

impl Estimator for RandomForestClassifier

impl Params for RandomForestClassifier

fn get_params(&self) -> Vec<(&'static str, ParamValue)>

Returns the current hyperparameters as (name, value) pairs.

fn set_params(&mut self, name: &str, value: ParamValue) -> Result<()>

Applies a named hyperparameter. Returns an error if name is unknown or value is the wrong type for that parameter.

impl PredictProba for RandomForestClassifier

fn predict_proba(&self, x: &Matrix) -> Result<Matrix>

Return per-class probability estimates.

impl Predictor for RandomForestClassifier

fn fit(&mut self, x: &Matrix, y: &[f64]) -> Result<()>

Fit the estimator on training features and target values.

fn predict(&self, x: &Matrix) -> Result<Vec<f64>>

Predict one numeric value per input row. For classifiers this is the predicted class label; probability estimates are exposed separately by PredictProba.

fn is_fitted(&self) -> bool

Whether the estimator has learned fitted state.

fn fit_predict(&mut self, x: &Matrix, y: &[f64]) -> Result<Vec<f64>>

Convenience: fit then predict on the same data.