I don't think there's a synonym file for this use case. I am not even sure if
synonym is the right way to handle it.

I think the better way to improve recall is to mark up your documents with
a "hidden" field of is the geographic relations. For example, before indexing,
you can add a field to all documents containing "South America", something
like: "South America is a subcontinent, that is consisted of the countries 
Brazil,
Chile, Argentina, …"

This data can come from various sources, such as wikipedia, wordnet, etc.


On Jul 5, 2012, at 4:12 AM, Stephen Lacy wrote:

> When a user types in South America they want to be able to see documents
> containing Brazil, Chile etc.
> No I have already thrown together a list of countries and continents
> however I'm a little more ambitious,
> I would like to get a lot more regions such as american states as well or
> Former members of the USSR...
> Are there ready made synonym files or taxonomies in a different format.
> Are synonyms the best way of achieving this? Perhaps there is a better way?
> Any pitfalls or advice on this subject from someone who has done this
> before would be appreciated.
> Thanks
> 
> Stephen

Reply via email to