String similarity

[Luca Montecchiani]

Tim Churches wrote:

 > This starts to look like a probabilistic record linkage (matching) problem. See the
 > Febrl project at http://datamining.anu.edu.au/projects/linkage.html - amongst
 > other things it contains a library of string comparators written in Python.

Thanks for the link, the stringcmp.py contains some cool code that I'll try later.
Unfortunally I can't gain speed but another bunch of algo to improve results quality ;)

ciao,
luca