Skip to main content

standardize_columns

matten_mlprep

Function standardize_columns 

Source 
pub fn standardize_columns(x: &Tensor) -> Result<Tensor, MattenMlprepError>
Expand description

Standardizes each column to zero mean and unit (population) standard deviation: out[i,j] = (x[i,j] - mean_j) / std_j.

std_j uses the population formula (divide by n), matching scikit-learn’s StandardScaler.

§Errors

use matten::Tensor;
use matten_mlprep::standardize_columns;

// Column 0: [1, 3] -> mean 2, std 1 -> [-1, 1]; column 1: [10, 20] -> [-1, 1].
let x = Tensor::new(vec![1.0, 10.0, 3.0, 20.0], &[2, 2]);
let z = standardize_columns(&x).unwrap();
assert_eq!(z.as_slice(), &[-1.0, -1.0, 1.0, 1.0]);
Examples found in repository?
examples/standardize_columns.rs (line 20)
17fn main() {
18    // 3 samples, 2 features.
19    let x = Tensor::new(vec![1.0, 10.0, 2.0, 20.0, 3.0, 30.0], &[3, 2]);
20    let z = standardize_columns(&x).expect("two non-constant columns");
21    println!("input  shape {:?}: {:?}", x.shape(), x.as_slice());
22    println!("z-score      {:?}: {:?}", z.shape(), z.as_slice());
23
24    // Shape is preserved and each standardized column has (near) zero mean.
25    assert_eq!(z.shape(), x.shape());
26    let zs = z.as_slice();
27    let col0_mean = (zs[0] + zs[2] + zs[4]) / 3.0;
28    let col1_mean = (zs[1] + zs[3] + zs[5]) / 3.0;
29    assert!(col0_mean.abs() < 1e-9 && col1_mean.abs() < 1e-9);
30    println!("standardize_columns: OK");
31}