0

Place names with common words

Hello,
I am creating an application that retrieves places from news articles using PlaceMaker. For the most part it works great, but the most common errors in identifying places deals with place names that use common words or phrases. For example:

Pine Bark, Waleska, GA, US
Lofty Heights, MD, US
German, NY, US
Basket Case, St. Augustine, FL, US
Greenhouse, Plant City, FL, US
Main Point, Newfoundland and Labrador, CA

The term "Chicago" only refers to a place, but the term "Main Point" could also be part of a discussion that has nothing to do with any place. I'm wondering if these types of place names could be identified and given a lower confidence because they could easilty be confused, especially if they are not a highly populated place? For now, I build a list of these places as I find them in the results and filter them from all future results. There are many more like this and it seems like they are never ending. I would be happy to send you my complete list. It is a work in progress.
Thanks,
Ken

by
1 Reply
  • QUOTE (KenB @ May 24 2011, 12:03 PM) <{POST_SNAPBACK}>
    Hello,
    I am creating an application that retrieves places from news articles using PlaceMaker. For the most part it works great, but the most common errors in identifying places deals with place names that use common words or phrases. For example:

    Pine Bark, Waleska, GA, US
    Lofty Heights, MD, US
    German, NY, US
    Basket Case, St. Augustine, FL, US
    Greenhouse, Plant City, FL, US
    Main Point, Newfoundland and Labrador, CA

    The term "Chicago" only refers to a place, but the term "Main Point" could also be part of a discussion that has nothing to do with any place. I'm wondering if these types of place names could be identified and given a lower confidence because they could easily be confused, especially if they are not a highly populated place? For now, I build a list of these places as I find them in the results and filter them from all future results. There are many more like this and it seems like they are never ending. I would be happy to send you my complete list. It is a work in progress.
    Thanks,
    Ken


    Placemaker tries to ignore phrases composed of common words like "Main" and "Point" where there is no location context. "Main Point" by itself should not trigger a match, but "Main Point, Newfoundland" should trigger a match. In some cases, our confidence may be too high and Placemaker will trigger a match when it shouldn't. Of the phrases you provided, only "Lofty Heights" triggered a match for me.

    It would help us to get a list of common phrases that wrongly return Placemaker results. It would be better if we could get the full sentence containing each common phrase so we can see the context for the usage.

    Eddie Babcock
    Yahoo! Geo Technologies
    0
This forum is locked.

Recent Posts

in Placemaker Enhancement Requests