In information theory and computer science, the Damerau–Levenshtein distance (named after Frederick J. Damerau and Vladimir I. Levenshtein) is a distance (string metric) between two strings, i.e., finite sequence of symbols, given by counting the minimum number of operations needed to transform one string into the other, where an operation is defined as an insertion, deletion, or substitution of a single character, or a transposition of two adjacent characters. In his seminal paper, Damerau not only distinguished these four edit operations but also stated that they correspond to more than 80% of all human misspellings. Damerau's paper considered only misspellings that could be corrected with at most one edit operation.
The Damerau–Levenshtein distance differs from the classical Levenshtein distance by including transpositions among its allowable operations. The classical Levenshtein distance only allows insertion, deletion, and substitution operations. Modifying this distance by including transpositions of adjacent symbols produces a different distance measure, known as the Damerau–Levenshtein distance.
While the original motivation was to measure distance between human misspellings to improve applications such as spell checkers, Damerau–Levenshtein distance has also seen uses in biology to measure the variation between protein sequences.
To express the Damerau–Levenshtein distance between two strings and a function is defined, whose value is a distance between an –symbol prefix (initial substring) of string and a –symbol prefix of .
The function is defined recursively as:
where is the indicator function equal to 0 when and equal to 1 otherwise.
Each recursive call matches one of the cases covered by the Damerau–Levenshtein distance:
- corresponds to a deletion (from a to b).
- corresponds to an insertion (from a to b).
- corresponds to a match or mismatch, depending on whether the respective symbols are the same.
- corresponds to a transposition between two successive symbols.
The Damerau–Levenshtein distance between a and b is then given by the function value for full strings: where denotes the length of string a and is the length of b.
Presented here are two algorithms: the first, simpler one, computes what is known as the optimal string alignment (sometimes called the restricted edit distance), while the second one computes the Damerau–Levenshtein distance with adjacent transpositions. Adding transpositions adds significant complexity. The difference between the two algorithms consists in that the optimal string alignment algorithm computes the number of edit operations needed to make the strings equal under the condition that no substring is edited more than once, whereas the second one presents no such restriction.
Take for example the edit distance between CA and ABC. The Damerau–Levenshtein distance LD(CA,ABC) = 2 because CA → AC → ABC, but the optimal string alignment distance OSA(CA,ABC) = 3 because if the operation CA → AC is used, it is not possible to use AC → ABC because that would require the substring to be edited more than once, which is not allowed in OSA, and therefore the shortest sequence of operations is CA → A → AB → ABC. Note that for the optimal string alignment distance, the triangle inequality does not hold: OSA(CA,AC) + OSA(AC,ABC) < OSA(CA,ABC), and so it is not a true metric.
Optimal string alignment distance
algorithm OSA-distance is input: strings a[1..length(a)], b[1..length(b)] output: distance, integer let d[0..length(a), 0..length(b)] be a 2-d array of integers, dimensions length(a)+1, length(b)+1 // note that d is zero-indexed, while a and b are one-indexed. for i := 0 to length(a) inclusive do d[i, 0] := i for j := 0 to length(b) inclusive do d[0, j] := j for i := 1 to length(a) inclusive do for j := 1 to length(b) inclusive do if a[i] = b[j] then cost := 0 else cost := 1 d[i, j] := minimum(d[i-1, j ] + 1, // deletion d[i, j-1] + 1, // insertion d[i-1, j-1] + cost) // substitution if i > 1 and j > 1 and a[i] = b[j-1] and a[i-1] = b[j] then d[i, j] := minimum(d[i, j], d[i-2, j-2] + cost) // transposition return d[length(a), length(b)]
The difference from the algorithm for Levenshtein distance is the addition of one recurrence:
if i > 1 and j > 1 and a[i] = b[j-1] and a[i-1] = b[j] then d[i, j] := minimum(d[i, j], d[i-2, j-2] + cost) // transposition
Distance with adjacent transpositions
The following algorithm computes the true Damerau–Levenshtein distance with adjacent transpositions; this algorithm requires as an additional parameter the size of the alphabet Σ, so that all entries of the arrays are in [0, |Σ|):
algorithm DL-distance is input: strings a[1..length(a)], b[1..length(b)] output: distance, integer da := new array of |Σ| integers for i := 1 to |Σ| inclusive do da[i] := 0 let d[−1..length(a), −1..length(b)] be a 2-d array of integers, dimensions length(a)+2, length(b)+2 // note that d has indices starting at −1, while a, b and da are one-indexed. maxdist := length(a) + length(b) d[−1, −1] := maxdist for i := 0 to length(a) inclusive do d[i, −1] := maxdist d[i, 0] := i for j := 0 to length(b) inclusive do d[−1, j] := maxdist d[0, j] := j for i := 1 to length(a) inclusive do db := 0 for j := 1 to length(b) inclusive do k := da[b[j]] ℓ := db if a[i] = b[j] then cost := 0 db := j else cost := 1 d[i, j] := minimum(d[i−1, j−1] + cost, //substitution d[i, j−1] + 1, //insertion d[i−1, j ] + 1, //deletion d[k−1, ℓ−1] + (i−k−1) + 1 + (j-ℓ−1)) //transposition da[a[i]] := i return d[length(a), length(b)]
To devise a proper algorithm to calculate unrestricted Damerau–Levenshtein distance note that there always exists an optimal sequence of edit operations, where once-transposed letters are never modified afterwards. (This holds as long as the cost of a transposition, , is at least the average of the cost of an insertion and deletion, i.e., .) Thus, we need to consider only two symmetric ways of modifying a substring more than once: (1) transpose letters and insert an arbitrary number of characters between them, or (2) delete a sequence of characters and transpose letters that become adjacent after deletion. The straightforward implementation of this idea gives an algorithm of cubic complexity: , where M and N are string lengths. Using the ideas of Lowrance and Wagner, this naive algorithm can be improved to be in the worst case.
Damerau–Levenshtein distance plays an important role in natural language processing. In natural languages, strings are short and the number of errors (misspellings) rarely exceeds 2. In such circumstances, restricted and real edit distance differ very rarely. Oommen and Loke even mitigated the limitation of the restricted edit distance by introducing generalized transpositions. Nevertheless, one must remember that the restricted edit distance usually does not satisfy the triangle inequality and, thus, cannot be used with metric trees.
Since DNA frequently undergoes insertions, deletions, substitutions, and transpositions, and each of these operations occurs on approximately the same timescale, the Damerau–Levenshtein distance is an appropriate metric of the variation between two strands of DNA. More common in DNA, protein, and other bioinformatics related alignment tasks is the use of closely related algorithms such as Needleman–Wunsch algorithm or Smith–Waterman algorithm.
The algorithm can be used with any set of words, like vendor names. Since entry is manual by nature there is a risk of entering a false vendor. A fraudster employee may enter one real vendor such as "Rich Heir Estate Services" versus a false vendor "Rich Hier State Services". The fraudster would then create a false bank account and have the company route checks to the real vendor and false vendor. The Damerau–Levenshtein algorithm will detect the transposed and dropped letter and bring attention of the items to a fraud examiner.
The U.S. Government uses the Damerau–Levenshtein distance with its Consolidated Screening List API.
- Brill, Eric; Moore, Robert C. (2000). An Improved Error Model for Noisy Channel Spelling Correction (PDF). Proceedings of the 38th Annual Meeting on Association for Computational Linguistics. pp. 286–293. doi:10.3115/1075218.1075255.
- Bard, Gregory V. (2007), "Spelling-error tolerant, order-independent pass-phrases via the Damerau–Levenshtein string-edit distance metric", Proceedings of the Fifth Australasian Symposium on ACSW Frontiers : 2007, Ballarat, Australia, January 30 - February 2, 2007, Conferences in Research and Practice in Information Technology 68, Darlinghurst, Australia: Australian Computer Society, Inc., pp. 117–124, ISBN 1-920-68285-6. The isbn produces two hits: a 2007 work and a 2010 work at World Cat.
- Li; et al. (2006). Exploring distributional similarity based models for query spelling correction (PDF). Proceedings of the 21st International Conference on Computational Linguistics and the 44th annual meeting of the Association for Computational Linguistics. pp. 1025–1032. doi:10.3115/1220175.1220304.
- Damerau, Fred J. (March 1964), "A technique for computer detection and correction of spelling errors", Communications of the ACM (ACM) 7 (3): 171–176, doi:10.1145/363958.363994
- Levenshtein, Vladimir I. (February 1966), "Binary codes capable of correcting deletions, insertions, and reversals", Soviet Physics Doklady 10 (8): 707–710
- The method used in: Majorek, Karolina A.; Dunin-Horkawicz, Stanisław; et al. (2013), "The RNase H-like superfamily: new members, comparative structural analysis and evolutionary classification", Nucleic Acids Research 42 (7): 4160–4179, doi:10.1093/nar/gkt1414
- Oommen, B. J.; Loke, R. K. S. "Pattern recognition of strings with substitutions, insertions, deletions and generalized transpositions". doi:10.1016/S0031-3203(96)00101-X. CiteSeerX: 10
.1 .1 .50 .1459.
- Lowrance, Roy; Wagner, Robert A. (April 1975), "An Extension of the String-to-String Correction Problem", JACM 22 (2): 177–183, doi:10.1145/321879.321880
- Navarro, Gonzalo (March 2001), "A guided tour to approximate string matching", ACM Computing Surveys 33 (1): 31–88, doi:10.1145/375360.375365