Note: This package supports Python 3.6 or newer.

This is a Python implementation of Continuous k-Nearest Neighbors(CkNN) proposed in the paper 'Consistent Manifold Representation for Topological Data Analysis' (https://arxiv.org/pdf/1606.02353.pdf)

This package only depends on numpy and scipy. The package can be installed via pip:

$ pip install git+https://github.com/chlorochrule/cknn

X is a data matrix. A simple example is like:

from cknn import cknneighbors_graph

ckng = cknneighbors_graph(X, n_neighbors=5, delta=1.0)

MIT