Pitney Bowes MapXtreme - Manual (Page 331)

Chapter 18: Geocoding
Understanding the Geocoding Model
MapXtreme v7.1 338 Developer Guide
GeocodingConstraints
This class contains the preferences that can be set when geocoding. All are set/get properties that
return true or false (default).
AddressCandidates
This class is a list of possible address matches that is returned in a GeocodeResponse.
BaseGeocodeMatchCode and GeocodeMatchCode
These are classes that implement the IGeocodeMatchCode interface. The interface exposes a
ResultCode property that is a string that describes how well the match was made.
BaseGeocodeMatchCode is returned when a request is sent to the Envinsa Location Utility Service.
GeocodeMatchCode extends BaseGeocodeMatchCode. It is returned when a geocode request is
sent to a MapMarker server. GeocodeMatchCode includes additional convenience properties to
determine which parts of the address matched. Among them, StreetAddressMatch and
MunicipalityMatch return a value of true if the candidate matched the street and municipality. For
details on result codes see
Understanding Accuracy for Close Matches .
CandidateAddress
This class defines an address that was geocoded. Its properties include Address, the address for the
geocoded address, GeocodeMatchCode, which explains how well the address was geocoded, and
Point, the geometry that represents the candidate address.
Understanding the Geocoding Model
The geocoding client is based on a model of relative matching that is governed by a set of weights
that scores each portion of the address against candidate records (possible matches) in the data.
The resulting scores are summed and the candidate’s total score is used to determine the best
match or matches. A close match is made when there is a candidate that scores well above other
candidates. In addition, the matching routine uses a set of geocoding preferences to determine
whether certain matching conditions are required or relaxed. For example, the default preferences
include relaxing a match on postal code, but requiring a match on the house number and street
name. This gives the best return of hit rate with the fewest erroneous matches (false positives) and
the best performance.
Geocoding Trade-offs
With a relative matching system such as the geocoding client, there are trade-offs that must be
considered in light of how you use the geocoded data. Consider questions like the following:
• What level of matching accuracy are you looking for (unique address match, close match)?
• What level of geographic accuracy do you need for your geocoded points (street level, ZIP Code
centroid)?