kmeans 2.0.2

Small and fast library for k-means clustering calculations.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33

use crate::api::DistanceFunction;
use crate::memory::*;
use crate::{KMeans, KMeansConfig, KMeansState};
use rand::prelude::*;
use std::simd::Simd;

#[inline(always)]
pub fn calculate<T, const LANES: usize, D>(kmean: &KMeans<T, LANES, D>, state: &mut KMeansState<T>, config: &KMeansConfig<'_, T>)
where
    T: Primitive,
    Simd<T, LANES>: SupportedSimdArray<T, LANES>,
    D: DistanceFunction<T, LANES>,
{
    let (assignments, centroids, centroid_frequency, k) =
        (&mut state.assignments, &mut state.centroids, &mut state.centroid_frequency, state.k);

    assignments.iter_mut().for_each(|a| {
        *a = config.rnd.borrow_mut().gen_range(0..k);
        centroid_frequency[*a] += 1;
    });
    kmean
        .p_samples
        .chunks_exact_stride()
        .zip(assignments.iter().cloned())
        .for_each(|(sample, assignment)| {
            centroids
                .bfr
                .iter_mut()
                .skip(centroids.stride * assignment)
                .zip(sample.iter().cloned())
                .for_each(|(cv, sv)| *cv += sv / T::from(centroid_frequency[assignment]).unwrap());
        });
}