clustering-tfjs/dist/utils/pairwise_distance.d.ts

High-performance TypeScript clustering algorithms (K-Means, Spectral, Agglomerative) with TensorFlow.js acceleration and scikit-learn compatibility

import * as tf from '../tf-adapter';
/**
 * Optimised Euclidean pairwise distance using the identity
 * ‖x − y‖² = ‖x‖² + ‖y‖² − 2·xᵀy to avoid building an (n,n,d) tensor.
 */
export declare function pairwiseEuclideanMatrix(points: tf.Tensor2D): tf.Tensor2D;
/**
 * Computes the pairwise distance matrix for the given points according to the
 * requested metric.
 *
 * The result is an `(n, n)` tensor `D` where `D[i, j]` contains the distance
 * between row `i` and row `j` of the input `points`.
 *
 * Supported metrics:
 *   • "euclidean"  – ℓ2 distance (uses an optimised implementation)
 *   • "manhattan"  – ℓ1 distance
 *   • "cosine"     – 1 − cosine-similarity
 *
 * For performance and numerical stability the computation is wrapped in
 * `tf.tidy` so that all intermediate tensors are eagerly disposed.
 */
export declare function pairwiseDistanceMatrix(points: tf.Tensor2D, metric?: 'euclidean' | 'manhattan' | 'cosine'): tf.Tensor2D;
//# sourceMappingURL=pairwise_distance.d.ts.map