pyFM.spectral.nearest_neighbor.brute_query

pyFM.spectral.nearest_neighbor.brute_query(X, Y, k=1, return_distance=False, n_jobs=None, working_memory=None, **_)

Nearest neighbours by brute force. Best above a handful of dimensions.

Scikit learn uses nice mixed precision with memory handling.

Parameters:

  • X (np.ndarray) – (n1, p). Reference points, the set being searched.

  • Y (np.ndarray) – (n2, p). Query points.

  • k (int) – Number of neighbours.

  • return_distance (bool) – Whether to also return distances.

  • n_jobs (int) – Passed through to scikit-learn.

  • working_memory (int, optional) – Chunking budget in MB. Defaults to the configured value.

Returns:

  • dists (np.ndarray, optional) – (n2,) if k = 1 else (n2, k). Distance to each neighbour. Only if return_distance.

  • matches (np.ndarray) – (n2,) if k = 1 else (n2, k). Index in X of each neighbour.