Struct FeatureImportance 

pub struct FeatureImportance {
    pub feature_name: String,
    pub importance_score: f64,
    pub standard_error: Option<f64>,
    pub confidence_interval: Option<(f64, f64)>,
    pub p_value: Option<f64>,
    pub stability: Option<f64>,
}

Feature importance from LIME

Fields

feature_name: String

Feature name

importance_score: f64

Importance score

standard_error: Option<f64>

Residual standard error of the feature’s local slope estimate, or None when it is not estimable (see Self::confidence_interval).

confidence_interval: Option<(f64, f64)>

95% confidence interval on the feature’s local slope, from the residual standard error of its univariate fit. None when the perturbations left too little variation in this feature (or too few samples) to estimate one. Previously the literal (coeff - 0.1, coeff + 0.1).

p_value: Option<f64>

Two-sided p-value for H0: slope = 0 on the same univariate fit, None under the same conditions. Previously the constant 0.05 for every feature of every instance.

stability: Option<f64>

Always None: measuring how stable a feature’s attribution is across perturbation rounds needs repeated independent LIME runs, which this analyzer does not perform. Previously the constant 0.8.

Trait Implementations

impl Clone for FeatureImportance

fn clone(&self) -> FeatureImportance

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for FeatureImportance

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for FeatureImportance

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for FeatureImportance

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

where __S: Serializer,

Serialize this value into the given Serde serializer.