Struct RegressionResults 

pub struct RegressionResults<F>
where
    F: Float + Debug + Display + 'static,
{

    pub coefficients: Array1<F>,
    pub std_errors: Array1<F>,
    pub t_values: Array1<F>,
    pub p_values: Array1<F>,
    pub conf_intervals: Array2<F>,
    pub r_squared: F,
    pub adj_r_squared: F,
    pub f_statistic: F,
    pub f_p_value: F,
    pub residual_std_error: F,
    pub df_residuals: usize,
    pub residuals: Array1<F>,
    pub fitted_values: Array1<F>,
    pub inlier_mask: Vec<bool>,
}

Structure to hold detailed regression results

Fields

coefficients: Array1<F>

Coefficients of the regression model

std_errors: Array1<F>

Standard errors for the coefficients

t_values: Array1<F>

t-statistics for each coefficient

p_values: Array1<F>

p-values for each coefficient

conf_intervals: Array2<F>

Confidence intervals for each coefficient (lower, upper)

r_squared: F

R-squared value (coefficient of determination)

adj_r_squared: F

Adjusted R-squared value

f_statistic: F

F-statistic for the regression

f_p_value: F

p-value for the F-statistic

residual_std_error: F

Residual standard error

df_residuals: usize

Degrees of freedom

residuals: Array1<F>

Residuals

fitted_values: Array1<F>

Fitted (predicted) values

inlier_mask: Vec<bool>

Boolean mask indicating inliers for robust methods like RANSAC This is only populated for methods that explicitly identify inliers/outliers

Implementations

impl<F> RegressionResults<F>

where F: Float + Debug + Display + 'static,

pub fn predict(&self, xnew: &ArrayView2<'_, F>) -> StatsResult<Array1<F>>

where F: Mul<Output = F> + Sum<F>,

Predict values using the regression model on new data.

Arguments

• x_new - New independent variables data (must have the same number of columns as the original x data)

Returns

Array of predicted values for each row in x_new.

Examples

use ndarray::{array, Array2};
use scirs2_stats::linear_regression;

// Fit a model
let x = Array2::from_shape_vec((3, 2), vec![
    1.0, 1.0,  // 3 observations with 2 variables (intercept and x1)
    1.0, 2.0,
    1.0, 3.0,
]).unwrap();
let y = array![3.0, 5.0, 7.0];  // y = 1 + 2*x1

let model = linear_regression(&x.view(), &y.view(), None).unwrap();

// Predict for new data
let x_new = Array2::from_shape_vec((2, 2), vec![
    1.0, 4.0,  // 2 new observations
    1.0, 5.0,
]).unwrap();

let predictions = model.predict(&x_new.view()).unwrap();

// Check predictions: y = 1 + 2*x1
assert!((predictions[0] - 9.0f64).abs() < 1e-8f64);  // 1 + 2*4 = 9
assert!((predictions[1] - 11.0f64).abs() < 1e-8f64); // 1 + 2*5 = 11

pub fn summary(&self) -> String

Return a summary of the regression results as a string.

Returns

A formatted string with regression statistics.