Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Artificially lower match scores for differing suffix directions & address systems #117

Open
Tracked by #447
stdavis opened this issue Aug 28, 2020 · 9 comments
Open
Tracked by #447
Labels
enhancement incomplete This issue needs more information to be actionable service - api geocoding the geocoding endpoint of the api

Comments

@stdavis
Copy link
Member

stdavis commented Aug 28, 2020

351 E 850 W, 84701 returns a match score of 92.43 but matches a totally different street: 351 E 850 N.

I would have expected the score to be much lower. I wonder if the API could artificially lower match scores when the suffix (or prefix) directions don't match the input.

We should also do this for candidates/matches that have a different address system.

Possible duplicate of #11

@steveoh steveoh added enhancement service - api geocoding the geocoding endpoint of the api labels Aug 31, 2020
@steveoh
Copy link
Member

steveoh commented Aug 31, 2020

It is certainly something we should investigate

@steveoh
Copy link
Member

steveoh commented Sep 7, 2022

I like this idea. We would need to figure out the value to lower the score by or if a default value should be used.

@stdavis
Copy link
Member Author

stdavis commented Sep 7, 2022

I'd like to start the bidding at -20 from the original score.

@steveoh
Copy link
Member

steveoh commented Sep 7, 2022

Typically that would put it at about 72. Would we want it to go below the default match threshold?

@stdavis
Copy link
Member Author

stdavis commented Sep 7, 2022

Good point. Is the default match threshold 75? Maybe -15 or -10?

@steveoh
Copy link
Member

steveoh commented Sep 7, 2022

its 70

@stdavis
Copy link
Member Author

stdavis commented Sep 7, 2022

Maybe -15 is the safest? Actually, though, I almost think that I would want it to drop below the match threshold. IMO, that's a bad match that only causes confusion. My vote would be maybe -25.

@stdavis
Copy link
Member Author

stdavis commented Jun 21, 2024

Should a similar thing be done for candidates that have mismatching address grids from the input address?

@steveoh
Copy link
Member

steveoh commented Jun 21, 2024

I like that idea

@stdavis stdavis changed the title Artificially lower match scores for differing suffix directions Artificially lower match scores for differing suffix directions & address systems Jun 21, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement incomplete This issue needs more information to be actionable service - api geocoding the geocoding endpoint of the api
Projects
None yet
Development

No branches or pull requests

2 participants