Struct StreamingShap 

pub struct StreamingShap { /* private fields */ }

Running mean of absolute SHAP values for online feature importance tracking.

Feed SHAP explanations as they are computed during training/prediction, and query the running importance scores at any time.

Example

let mut tracker = StreamingShap::new(n_features);
for sample in stream {
    let shap = model.explain(&sample.features);
    tracker.update(&shap);
}
let importances = tracker.importances(); // running mean |SHAP|

Implementations

impl StreamingShap

pub fn new(n_features: usize) -> Self

Create a new tracker for the given number of features.

pub fn update(&mut self, shap: &ShapValues)

Update with a new SHAP explanation, incrementing the running mean.

pub fn importances(&self) -> &[f64]

Current running mean of |SHAP| per feature.

pub fn count(&self) -> u64

Number of explanations processed.

pub fn reset(&mut self)

Reset to zero state.

Trait Implementations

impl Clone for StreamingShap

fn clone(&self) -> StreamingShap

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for StreamingShap

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

Formats the value using the given formatter.