Function kmedoids::fasterpam

pub fn fasterpam<N: NumAssignOps + Signed + Zero + PartialOrd + Copy + SafeAdd>(
    mat: &Array2<N>, 
    med: &mut Vec<usize>, 
    maxiter: usize
) -> (N, usize, usize, Vec<usize>)

Run the FasterPAM algorithm.

Given a dissimilarity matrix of size n x n, use:

let mut rng = rand::thread_rng();
let mut meds = random_initialization(n, k, &mut rng);
let (loss, numswap, numiter, assignment) = fasterpam(data, &mut meds, 100);

• type N - some signed data type such as i32 or f64
• mat - a pairwise distance matrix
• med - the list of medoids
• maxiter - the maximum number of iterations allowed

returns a tuple containing:

• the final loss
• the number of swaps performed
• the number of iterations needed
• the final cluster assignment