- Joined
- Jul 10, 2007
- Messages
- 6,306
Let's say that we have a group of points in 2D space. The group should be split up into clusters of 4 as efficiently as possible (maximize cluster sizes). A point can only be part of a cluster if the cluster size after the dot is added is less than or equal to the max cluster size (circle at top right).
Any ideas?
I was thinking of calculating center points for sets of data and then splitting it up using binary partitioning on the center points until set <= 4, but that wouldn't split it up into maximum cluster sizes. It'd split it up into even cluster sizes.
So... dunno what to do : (.
Below is a picture of me connecting dots for sets of 4 points to represent clusters. I need an algorithm to do that in code.
Any ideas?
I was thinking of calculating center points for sets of data and then splitting it up using binary partitioning on the center points until set <= 4, but that wouldn't split it up into maximum cluster sizes. It'd split it up into even cluster sizes.
So... dunno what to do : (.
Below is a picture of me connecting dots for sets of 4 points to represent clusters. I need an algorithm to do that in code.
