Hi Robert Perhaps not surprisingly I see this as an analog of of a common problem in bioinformatics, where you want to know if stretches of DNA are related.
this is a dynamic programming problem with penalties for edits the solution with the smallest edit distance is your goal. The penalties for particular edits can be empirically determined if you have a large set that has been solved or you could come up with a with a scoring matrix wit the penalty for one symbol to become another=20 (or a gap/insertion). "basic local alignment search tool" BLAST , "smith waterman algorithm" are some search terms that come to mind Robert M. M=FCnch wrote: > Hi all, yesterday I had a situation where I had to hunt a difference =20 > between some numbers and I thought if this problem can be solved in =20 > general. Here is the situation: >=20 > You have one sum: 41695,83 >=20 > And a set of numbers: 6594,14 + 981,75 + 8747,39 + 1457,90 + 15114,65 >=20 > The sum you have doesn't match the sum of the single numbers. >=20 > Now the question is: What (minimal) typos in the set of numbers leads t= o =20 > the sum you have? >=20 > I find this an interesting question. If it's possible to derive a set o= f =20 > possible typos these sets could be ranked by some propability. And than= it =20 > would be possible to check differences in data sets back to typos. >=20 > I hope it's clear what I mean. So, has anyone any idea if this is =20 > solvable? Robert -- To unsubscribe from the list, just send an email to lists at rebol.com with unsubscribe as the subject.
