Nearest-centroid assignment via fused squared-distance computation

Computes \(D[i,k] = \|x_i\|^2 - 2 x_i^\top c_k + \|c_k\|^2\) and returns \(\arg\min_k D[i,k]\) for each row \(i\), 1-indexed. GPU path avoids host round-trips by chaining resident operations.

Usage

pairwise_sqdist_argmin(X, Ct, x_norms = NULL, c_norms = NULL)

Arguments

X

n×p adgeMatrix or plain matrix (query points).

Ct

p×K numeric matrix holding the centroids transposed: nrow(Ct) is the feature dimension \(p\) and each column is a centroid. Pass t(centroids) when your centroids are stored k×p with one centroid per row. A dimension mismatch (ncol(X) != nrow(Ct)) errors, but a square k == p matrix passed untransposed cannot be detected and will silently assign to the wrong (column) centroids.

x_norms

Optional n-vector of precomputed \(\|x_i\|^2\). Computed if NULL.

c_norms

Optional K-vector of precomputed \(\|c_k\|^2\). Computed if NULL.

Value

Integer vector of length n, 1-indexed nearest centroid per row.