1

I have a requirement to found the fraud customer's , the customer name which are all matching 80% with the fraudulent customer's name. So any one give idea for pattern check. I think this explanation is sufficient to understand.

Sathyajith Bhat
  • 21,321
  • 22
  • 95
  • 134
Nagarajan
  • 432
  • 3
  • 12
  • 28

3 Answers3

3

There are set of standard algorithmns which will help you to acheive the requirement like ..

  1. Levenstein
  2. JaroWinkler
  3. NeedlemanWunch
  4. SmithWaterman
  5. ChapmanLengthDeviation..
Ashley John
  • 2,379
  • 2
  • 21
  • 36
  • Thanks friends ... really these concepts are new to my life... thanks fr a valuable suggetion.. i will try this now.. – Nagarajan Nov 21 '11 at 10:14
1

You can use Levenshtein for this see Levenshtein distance in T-SQL

The example is in T-sql but you can use it to see the pattern

Community
  • 1
  • 1
Ivo
  • 3,406
  • 4
  • 33
  • 56
0

I just created a SQLCLR library for running Levenshtein on Sql Server here. Feel free to contribute. :)

Robert Jeppesen
  • 7,837
  • 3
  • 35
  • 50