algorithm - What are some fast approximations of Nearest Neighbor? -
say have huge (a few million) list of n vectors, given new vector, need find pretty close 1 set doesn't need closest. (nearest neighbor finds closest , runs in n time)
what algorithms there can approximate nearest neighbor @ cost of accuracy?
edit: since help, should mention data pretty smooth of time, small chance of spikiness in random dimension.
there exist faster algoritms o(n) search closest element arbitary distance. check http://en.wikipedia.org/wiki/kd-tree details.
Comments
Post a Comment