Struct Influence 

pub struct Influence {
    pub hat_diag: Array1<f64>,
    pub resid_studentized_internal: Array1<f64>,
    pub resid_studentized_external: Array1<f64>,
    pub cooks_distance: Array1<f64>,
    pub dffits: Array1<f64>,
}

Per-observation OLS influence diagnostics.

Every field is an n-vector aligned with the rows of the design matrix.

Fields

hat_diag: Array1<f64>

Hat-matrix diagonal (leverage), h_i = x_i' (X'X)^{-1} x_i.

resid_studentized_internal: Array1<f64>

Internally studentized residuals.

resid_studentized_external: Array1<f64>

Externally (leave-one-out) studentized residuals.

cooks_distance: Array1<f64>

Cook’s distance.

dffits: Array1<f64>

DFFITS.

Implementations

impl Influence

pub fn new(res: &LinearResults, exog: &Array2<f64>) -> Influence

Compute the influence diagnostics from a fitted LinearResults and the design matrix exog that produced it.

exog must be the same n × k design used in the fit (including any constant column). The hat diagonal is formed from the model’s normalized_cov_params, which for OLS equals (X'X)^{-1}.

Panics

Panics if exog’s row count does not match the number of residuals.

Trait Implementations

impl Clone for Influence

fn clone(&self) -> Influence

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Influence

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

Formats the value using the given formatter.