Struct HypothesisTest 

pub struct HypothesisTest;

Hypothesis testing methods

Implementations

impl HypothesisTest

pub fn wald_test( beta: &Array1<f64>, cov_matrix: &Array2<f64>, r: &Array2<f64>, q: &Array1<f64>, ) -> Result<(f64, f64, usize), GreenersError>

Wald test for linear restrictions on coefficients

Tests H₀: R·β = q against H₁: R·β ≠ q

Arguments

• beta - Coefficient estimates (k × 1)
• cov_matrix - Variance-covariance matrix (k × k)
• r - Restriction matrix (m × k) where m = number of restrictions
• q - Restriction values (m × 1), usually zeros

Returns

Tuple of (wald_statistic, p_value, degrees_of_freedom)

Examples

use greeners::HypothesisTest;
use ndarray::{Array1, Array2};

// Setup de dados falsos para o exemplo
let beta = Array1::from(vec![0.5, 2.0, 0.0]);
let cov_matrix = Array2::eye(3); // Matriz identidade como exemplo

// Testando se beta[1] = 0 e beta[2] = 0
let r = Array2::from_shape_vec((2, 3), vec![
    0.0, 1.0, 0.0,
    0.0, 0.0, 1.0,
])?;
let q = Array1::from(vec![0.0, 0.0]);

let (wald_stat, p_value, df) = HypothesisTest::wald_test(&beta, &cov_matrix, &r, &q)?;

pub fn f_test_nested( ssr_restricted: f64, ssr_full: f64, n: usize, k_full: usize, k_restricted: usize, ) -> Result<(f64, f64, usize, usize), GreenersError>

F-test for nested models (OLS specific)

Tests whether restricted model is adequate vs full model

Arguments

• ssr_restricted - Sum of squared residuals from restricted model
• ssr_full - Sum of squared residuals from full model
• n - Number of observations
• k_full - Number of parameters in full model
• k_restricted - Number of parameters in restricted model

Returns

Tuple of (f_statistic, p_value, df_numerator, df_denominator)

Formula

F = [(SSR_r - SSR_f) / (k_f - k_r)] / [SSR_f / (n - k_f)]

pub fn joint_significance( beta: &Array1<f64>, cov_matrix: &Array2<f64>, has_intercept: bool, ) -> Result<(f64, f64, usize), GreenersError>

Joint significance test (all coefficients except intercept = 0)

Convenience wrapper for Wald test of all slope coefficients

Arguments

• beta - Coefficient estimates (including intercept)
• cov_matrix - Variance-covariance matrix
• has_intercept - Whether first coefficient is intercept

Returns

Tuple of (test_statistic, p_value, degrees_of_freedom)