MinimalKNN: minimal package to construct k-NN Graph
Project description
A Minimal k-Nearest Neighbor Graph Construction Library
Overview
This package will provide a function to construct an approximated k-Nearest Neighbor graph from a list of three dimensional points. The graph construction algorithm is based on NN-descent presented in Dong, Moses, & Li (2011)[^DML2011]. The Euclidean and Manhattan metrics are implemented in the current version, while only the Euclidean one is available in Python. The algorithm efficiently constructs an approximated k-Nearest Neighbor graph. This provides a portable C++11 header and a Python interface.
Dependencies
The library is written in C++11 and do not depends on any library outside of the STL
. The Python interface is depends on NumPy
, and functional test procedures depend on Matplotlib
. The library is developed on g++
version 5.4 installed in Linux Mint 18.1 (serena). The Python interface is developed on Python 3.7.1 and Numpy 1.18.1.
References
[^DML2011]: Wei Dong, Charikar Moses, & Kai Li, WWW'11: Proceedings of the 20th international conference on World wide web (2011), 577--586 (doi: 10.1145/1963405.1963487)
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for minimalKNN-0.7-cp39-cp39-manylinux1_x86_64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | a3437da478ba1c09c776e3e6f853e607aa54280be28b4bfd2b2616889b758344 |
|
MD5 | 73f182e29167fa05b6d9dee0b3ab9ee5 |
|
BLAKE2b-256 | b618f7045e48a5e2961e60d9145f0f898ab021568bf1f4274526e52bf67c7404 |