7

I have a set of strings $S$ and I am using the edit-distance (Levenshtein) to measure the distance between all pairs.

Is there an algorithm for finding the string $x$ which minimizes the sum of the distances to all strings in $S$, that is

$\arg_x \min \sum_{s \in S} \text{edit-distance}(x,s)$

It seems like there should, but I can't find the right reference.

Raphael
  • 73,212
  • 30
  • 182
  • 400
Jose M Vidal
  • 173
  • 3

1 Answers1

5

The problem is known as "median string problem" and it is NP-complete; some results can be found searching with Google; in particular "2-Approximation Algorithms for Median and Centre String Problems". If $x$ must be one of the points in $S$ then the problem becomes solvable in polynomial time.

Vor
  • 12,743
  • 1
  • 31
  • 62