Distinguish between full target match and partial target match

[agarzola]

Hi!

I’m using natural to identify bus stops within text delivered via SMS and I’ve got the following situation: When a user message says “eastgate town center”, the Bayes Classifier gives the following two stop entries equal scoring:

• EASTGATE TOWN CENTER
• EASTGATE TOWN CENTER LOOP + CONVERGYS

Both of those get 0.0016474464579901153, which is the highest score. Is there any way to have natural distinguish between the above?

Thanks!

[kkoch986]

Sorry havent forgotten about this one will try to get to it asap. Haven't used the classifier in awhile so i'll have to dig around and refresh my memory.

[agarzola]

Thanks! We ended up having to do an overhaul of bus stop names and found that the second bus stop described above was misnamed. Likewise, a couple of others with similar issue were more accurately renamed such that this is no longer an issue. By all means, if you find a solution I’d love to hear it (I have future projects I’d like to use natural for and may run into this again), but there is no need to rush on it by any means.

Thanks again!