An optimal randomized algorithm for maximum Tukey depth

  • Authors:
  • Timothy M. Chan

  • Affiliations:
  • Univ. of Waterloo, Waterloo, Ontario, Canada

  • Venue:
  • SODA '04 Proceedings of the fifteenth annual ACM-SIAM symposium on Discrete algorithms
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

We present the first optimal algorithm to compute the maximum Tukey depth (also known as location or halfspace depth) for a non-degenerate point set in the plane. The algorithm is randomized and requires O(n log n) expected time for n data points. In a higher fiexed dimension d ≥ 3, the expected time bound is O(nd-1), which is probably optimal as well. The result is obtained using an interesting variant of the author's randomized optimization technique, capable of solving "implicit" linear-programming-type problems; some other applications of this technique are briefly mentioned.