Sovereign 8s.is mirror of standardgalactic/benchmark_spaces
Find a file
Repository files (latest commit first)
Filename Latest commit message Latest commit date
2023-03-05 22:26:40 -08:00
benchmark.py Updated commentary and added visualizations 2023-03-05 22:07:50 -08:00
Commentary.txt Updated commentary and added visualizations 2023-03-05 22:07:50 -08:00
README.md improved explanation 2023-03-05 22:20:25 -08:00
sphere_vs_projective_sphere.png added rendering of graph 2023-03-05 22:26:40 -08:00

benchmark_spaces

Implements different two dimensional spaces for clustering algorithms

Contains code for making animations of simulated annealing of points repelling each other and also for measuring the average deviation in distances between random points (lower is better).

The upshort of this is:

If you're doing two dimensional visualtizations you should use the geometry called 'twisted torus'.

Otherwise if having exact opposites isn't part of expected/wanted semantics of your data you should use Projective Sphere.

In the unusual case of polar opposites being expected you should use Sphere.

More detailed thoughts are in Commentary.txt