variance_simd

scirs2_core::ndarray_ext::elementwise

Function variance_simd 

Source 
pub fn variance_simd<F>(a: &ArrayView1<'_, F>) -> F
where
    F: Float + SimdUnifiedOps,
Expand description

SIMD-accelerated variance computation

Computes the sample variance: Var(x) = sum((x - mean)²) / (n-1) Uses Bessel’s correction for unbiased estimation.

§Arguments

  • a - Input array

§Returns

Sample variance of the array

§Examples

use scirs2_core::ndarray::array;
use scirs2_core::ndarray_ext::elementwise::variance_simd;

let x = array![2.0_f64, 4.0, 4.0, 4.0, 5.0, 5.0, 7.0, 9.0];
let var = variance_simd::<f64>(&x.view());
// Sample variance: sum of squared deviations = 32, n = 8
// var = 32 / (8-1) = 32/7 ≈ 4.571
let expected = 32.0 / 7.0;
assert!((var - expected).abs() < 1e-10);

§Use Cases

  • Statistical analysis
  • Quality control (process capability)
  • Risk assessment (financial variance)
  • Feature scaling (standardization)