Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.programming > #3235

Re: document classification problem?

Show all headers | View raw

On Friday, March 29, 2013 3:40:02 PM UTC+5:30, Arun Jayapal wrote:
> Hi,
> 
> 
> 
> I have an interesting problem. It goes like this:
> 
> 
> 
> I have a list of "ideal" product names; each product is linked to some code (say class A, B, C etc. You can imagine anything for a product name like for e.g. a csv file of such a file would look like the follows:
> 
> 
> 
> ---
> 
> ItemName,Class
> 
> Philips Plasma LED TV - Model 45321Q,A
> 
> Angle Bracket 1.0" x 2.25" x 4.0",B
> 
> Spray Paint - Black 150 ml,C
> 
> ...
> 
> .
> 
> .
> 
> ---
> 
> 
> 
> Please visualize the above as a table. The above file is what I call the master file. It can contain millions of records.
> 
> 
> 
> Next I am given a dump of invoices which contain item names. The goal is to apply "class" to each invoice line item description. 
> 
> 
> 
> However the data you see an invoice line item isn't exactly what you'd get in the master file. It can contain so many other attributes describing the item. To explain the scenario, take the case of a simple pencil:
> 
> 
> 
> The entry in the master file may go like:
> 
> 
> 
> ---
> 
> 
> 
> ItemName,Class
> 
> ...
> 
> ...
> 
> HB Lead Pencil,C
> 
> ...
> 
> ...
> 
> 
> 
> ---
> 
> 
> 
> But it the transaction file the corresponding item can appear as:
> 
> 
> 
> 20Nos Stadler 1.5 HB Lead Pencil
> 
> 
> 
> 
> 
> It can appear to be jumbled up in any order. Hence there will never be a direct match. But still a human can gets hints and probably classify that item with class 'C'.
> 
> 
> 
> Given all these problems how should you write a computer program to solve it?
> 
> 
> 
> -arun

Back to comp.programming | Previous | Next — Previous in thread | Find similar | Unroll thread

Thread

document classification problem? Arun Jayapal <deostroll@gmail.com> - 2013-03-29 03:10 -0700
  Re: document classification problem? "Chris Uppal" <chris.uppal@metagnostic.REMOVE-THIS.org> - 2013-03-29 12:44 +0000
    Re: document classification problem? Arun Jayapal <deostroll@gmail.com> - 2013-04-09 12:06 -0700
  Re: document classification problem? Ben Bacarisse <ben.usenet@bsb.me.uk> - 2013-03-29 14:48 +0000
  Re: document classification problem? Torsten Eichstädt <torsten.eichstaedt@FernUni-Hagen.de> - 2013-03-30 12:51 +0100
    Re: document classification problem? "Chris Uppal" <chris.uppal@metagnostic.REMOVE-THIS.org> - 2013-03-30 16:15 +0000
      Re: document classification problem? Torsten Eichstädt <torsten.eichstaedt@FernUni-Hagen.de> - 2013-03-31 17:27 +0200
        Re: document classification problem? "Chris Uppal" <chris.uppal@metagnostic.REMOVE-THIS.org> - 2013-04-06 14:06 +0100
  Re: document classification problem? Arun Jayapal <deostroll@gmail.com> - 2013-04-09 12:05 -0700

csiph-web