If I were to construct a KD tree representing a sufficiently dense space, could it be used as a decision tree by querying it for the nearest neighbor to the point to classify?
Add a comment |
1 Answer
$\begingroup$ $\endgroup$
Yes, you can. The scheme you are thinking of is called the nearest-neighbors classifier. Read the link for more details. In many settings it turns out to be a quite effective classification algorithm.
