Struct IsolationForest 

pub struct IsolationForest { /* private fields */ }

Isolation Forest for anomaly detection.

Uses an ensemble of isolation trees to detect outliers based on path length. Anomalies are easier to isolate (shorter paths) than normal points.

Algorithm

1. Build N isolation trees on random subsamples
2. Each tree recursively splits data by random feature + random threshold
3. Compute average path length across all trees
4. Convert to anomaly score (shorter path = more anomalous)
5. Use contamination parameter to set classification threshold

Examples

use aprender::prelude::*;

let data = Matrix::from_vec(
    6,
    2,
    vec![
        2.0, 2.0, 2.1, 2.0, 1.9, 2.1, 2.0, 1.9,  // Normal
        10.0, 10.0, -10.0, -10.0,                 // Outliers
    ],
)
.expect("Valid matrix dimensions and data length");

let mut iforest = IsolationForest::new()
    .with_contamination(0.3)
    .with_random_state(42);
iforest.fit(&data).expect("Fit succeeds with valid data");

// Predict returns 1 for normal, -1 for anomaly
let predictions = iforest.predict(&data);

// score_samples returns anomaly scores (lower = more anomalous)
let scores = iforest.score_samples(&data);

Performance

• Time complexity: O(n log m) where n=samples, m=max_samples
• Space complexity: O(t * m) where t=n_estimators

Implementations

impl IsolationForest

pub fn new() -> Self

Create a new Isolation Forest with default parameters.

Default: 100 trees, auto max_samples, 0.1 contamination

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

Set the number of trees in the ensemble.

pub fn with_max_samples(self, max_samples: usize) -> Self

Set the number of samples to draw for each tree.

pub fn with_contamination(self, contamination: f32) -> Self

Set the expected proportion of anomalies (0 to 0.5).

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

Set random seed for reproducibility.

pub fn is_fitted(&self) -> bool

Check if model has been fitted.

pub fn fit(&mut self, x: &Matrix<f32>) -> Result<()>

Fit the Isolation Forest on training data.

pub fn score_samples(&self, x: &Matrix<f32>) -> Vec<f32>

Compute anomaly scores for samples.

Returns a vector of scores where lower scores indicate higher anomaly likelihood.

pub fn predict(&self, x: &Matrix<f32>) -> Vec<i32>

Predict anomaly labels for samples.

Returns 1 for normal points and -1 for anomalies.

Trait Implementations

impl Clone for IsolationForest

fn clone(&self) -> IsolationForest

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for IsolationForest

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

Formats the value using the given formatter.

impl Default for IsolationForest

fn default() -> Self

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for IsolationForest

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for IsolationForest

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.