toha-nearest-neighbor 0.2.0

No project description provided

toha_nearest_neighbor

Serial and parallel bindings to brute force methods for 2D nearest neighbor problems in python

documentation

Benchmarks

Some basic benchmarks have been carried out against sklearn.neighbors to show the relative performance of this library. These preliminary results show better performance and better scaling in every function. However, keep in mind that sklearn handles a generic n-dimensional space while this package has been simplified to work with 2D data. Moreover, sklearn also has an additional algorithm ball_tree that scales to higher dimensions (N > 15) much better than kd-trees.