Nearest Neighours and k-d tree
==============================

Distances in two dimensions or three dimensions
are quite common and we sometimes assumes distances
in high dimension are friendly too. So why does a simple
model like the k-nearest neighbours seem to have difficulties
to predict ?

Problem definition
++++++++++++++++++

The first notebooks describes the problem.

.. toctree::
    :maxdepth: 2

    ../notebooks/nearest_neighbours_sparse_features