Approximations and Optimal Geometric Divide-And-Conquer

We give an efficient deterministic algorithm computing .5-approximations for range spaces of bounded VC-dimension. We assume that a npoint range space Z = (X, 72) of VC-dimension d is given to us by an oracle, which given a subset A G X, returns a list of all distinct sets of the form Anl?; R e %?(in time O(]Alk), where k is the number of sets returned… CONTINUE READING