Skip to main content

correlation

linreg_core::stats

Function correlation 

Source 
pub fn correlation(x: &[f64], y: &[f64]) -> f64
Expand description

Calculates the correlation coefficient (Pearson’s r) between two slices.

This implementation uses a numerically stable single-pass algorithm that avoids catastrophic cancellation, similar to Welford’s method. It computes mean, variance, and covariance in one pass.

§Arguments

  • x - First slice of f64 values
  • y - Second slice of f64 values (must be same length as x)

§Returns

The correlation coefficient as f64, or NaN if inputs are invalid

§Examples

use linreg_core::stats::correlation;

let x = vec![1.0, 2.0, 3.0, 4.0, 5.0];
let y = vec![2.0, 4.0, 5.0, 4.0, 5.0];
let r = correlation(&x, &y);
assert!((r - 0.7746).abs() < 1e-4);