Performance?
The highlighted comment was created in this revision.
I'm curious, have you measured the performance of this r-tree? During my extensive nearest neighbor search experiments in the past, I did make an attempt at an r-tree, but I found at least my impelementation to provide far inferior performance to my bucket kd-tree.
Yes. Now this tree in battles against Diamond in average takes 56,800 ns for RS. I know that it's not completly correctly to compare them, but yours kD-Tree do kNN with same k in about 70,000-80,000 ns. RS & kNN are performed more than 3000 times with 30 trees with 1-5 dimensions and 0- ~3200 points.
I plan to try use RS with dinamically calculate hypercube side in gun, which, i hope will produce pretty equal results with kNN, and in this case i will publish comparsion of RS & kNN for gun
