Module 2: Distances
  Lecture 8: Earth Mover's Distance (EMD)
 

                                            

 

 

Edit distance
  • Also known as Levenshtein distance
  • Used for words and strings
  • Minimum number of edit operations required to transform word to word
  • Edit operations on characters are
 
  • Insertion
 
  • Deletion
 
  • Substitution
  • Insertions and deletions can be treated as substitutions using gap
  • Example: each edit operation costs 1

 

idea

                        

deal

idea-

 

-deal

 
  • Edit distance is 2
  • Metric?