Re: Determining the similarity between a user supplied short piece of text (between 5 and 15 characters) and a list of similar length text items.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Thu, 2010-07-15 at 16:54 +0100, Richard Quadling wrote:

> Hi.
> 
> It seems that users cannot enter a vehicle registration 100% accurately.
> 
> We have recently released a small mobile web app which allows service
> engineers/inspectors to enter a vehicle registration number and a pin
> number to get service history for the vehicle.
> 
> We are getting around a 40% fail rate on the registrations for the
> first time of entry. This drops to around a 1.5% error rate on the
> second attempt.
> 
> Most of the time it is simply a case of 2 letters/numbers being
> swapped. Sometimes a letter/number is entered for a number/letter.
> 
> 0/O
> 1/I/l
> 2/Z
> 3/E
> 4/A
> 5/S
> 6/G
> 7/T
> 8/B
> 9/q
> 
> Some of the registrations are private and don't obey any format (NOTE:
> Trailer registrations aren't the same as vehicle registrations - they
> can be anything the owner wants. In some cases we have them as
> straight numbers - 1, 2, 3, etc.).
> 
> I'm looking for is a way to compare what they've entered against a
> known list and to provide my 10 best guesses.
> 
> What I'm stuck on is what criteria do I use.
> 
> I think something like the old style colour Mastermind game (right
> colour in the right place, right colour wrong place, wrong colour).
> But that's going to be slow. One of the contracts has over 30,000
> vehicles/trailers available to them.
> 
> Any suggestions really.
> 
> Regards,
> 
> Richard.
> 


I'm not sure how well it would work for things like a registration
number, but metaphone keys are pretty good for detecting similar words
based on phonetic sounds. If you don't get a match first time, maybe
convert all the digits in the reg no. to letters, and create a metaphone
key that you compare against the list of other keys in your DB that you
already created in the same manner (you'll need to create these first
time)

Like I said, I'm not sure how well it will work, but it might possibly
reduce the failure rate a bit.

Also, there's the electric shock treatment. Find out which engineers are
the worst typists, and... well you get the idea!

Thanks,
Ash
http://www.ashleysheridan.co.uk



[Index of Archives]     [PHP Home]     [Apache Users]     [PHP on Windows]     [Kernel Newbies]     [PHP Install]     [PHP Classes]     [Pear]     [Postgresql]     [Postgresql PHP]     [PHP on Windows]     [PHP Database Programming]     [PHP SOAP]

  Powered by Linux