Select to view content in your preferred language

Geocoding Score Documentation:  How is the score value determined?

12936
15
08-16-2012 11:26 AM
NathanLowry
Deactivated User
I know geocoding is largely controled by the Locators .xml files in C:\Program Files\ArcGIS\Desktop10.0\Locators.  However, where is it written (documented in white papers, technical articles, etc.), how the geocoding Score value is calculated?

I'd like to reference the documentation in metadata.

Thanks,

Nathan Lowry, GISP

Governor's Office of Information Technology
GIS Outreach Coordinator, State of Colorado
601 East 18th Avenue, Suite 220, Denver, CO 80203-1494
303.764.7801 720.402.4462 Cell nathan.lowry@state.co.us, http://www.colorado.gov/oit

How am I doing?  Please contact my manager Jon Gottsegen (Jon.Gottsegen@state.co.us) for comments or questions.
Tags (2)
0 Kudos
15 Replies
BradNiemand
Esri Regular Contributor

Jian,

1. The online service isn't a "True" composite locator.  It works differently and has additional functionality in it that can work this way.

2. We are actually changing the Minimum Match Score for the online service to be 93 as well for the USA and because your just using batch and Minimum Candidate Score is only used in FindAddressCandidates this is a pretty safe change that will actually reduce potential false positive matches.

3. This option is much higher risk than #2 above.

Brad

0 Kudos
JianLiu
Occasional Contributor

Brad Niemand

Brad,

Thanks for the info! I will change the Minimum Candidate and Minimum Match Scores for AddressPoint to 93 to test. Thanks!

A side question, but still kind of related to the geocoding algorithm ---

Our current geocoding procedure is

1. address standardization -- prep clean the addresses, including removal of confusing or redundant components such as suite numbers or p.o. boxes

2. address "cleansing" using 3rd party software such as QAS (I don't quite know what's in their box)

3. feed the "cleaned" address to ESRI geocoder

I have noticed there are various alias name settings in the .loc.xml files, so I assume the geocoder does address standardization as a prep step before doing actual address matching. So I was wondering if some of our procedures might be redundant? In all, I want to know what you would recommend as a cleansing step, in order to get the best out of the geocoder? Thank you.

regards,

Jian

0 Kudos
BradNiemand
Esri Regular Contributor

Jian,

Although we do handle the removal of confusing or redundant components while matching you may see a slightly better (1-2%) geocoding result if you use the standardization and "cleansing" process before.  It really depends on how bad the input address data is first and how good the standardization and "cleansing" tools are.  If I were you, I might give it a try with and without and see what the difference is.  You might see that you get little or even no better results.

Brad

0 Kudos
JianLiu
Occasional Contributor

Thanks Brad. We've done various comparisons and the results are not self explanatory themselves. Sometimes it could be better and sometimes worse. You are right that it depends on how good the cleansing tool is, and unfortunately I am not confident QAS is a good tool.

If anyone sees this post, I would appreciate it if someone can recommend a good address cleansing tool. Thanks in advance.

Also Brad, maybe I direct your attention to my other post here about geocoder again:

new geocoder is so much slower than the old one

Thanks for the information and suggestions!

Regards,

Jian 

0 Kudos
NathanLowry
Deactivated User

Similarly, one of our agencies was frustrated with the same thing.  There doesn't seem to be an option to allow the candidate among individual address locators with the highest score to be automatically selected as the match in a composite address locator - or, at least I couldn't find that option in any documentation or discussion thread yet.  So, last week I created Composite Address Locators by Score in ArcGIS Ideas.

Is such modification possible, and if so, how would it be done?

Sincerely, Nathan Lowry

GIS Outreach Coordinator

P 303.764.7801 |  F 303.764.7764

601 East 18th Avenue, Suite 220, Desk D-23, Denver, CO 80203-1494

nathan.lowry@state.co.us  |  www.colorado.gov/oit

How am I doing?  Please contact my manager Jon Gottsegen (jon.gottsegen@state.co.us) for comments or questions.

0 Kudos
NathanLowry
Deactivated User

Be glad to open it up in another post, if most appropriate.

0 Kudos