Introduction to the Validate Address Confidence Algorithm
ValidateAddress computes a confidence score for each validated address. This score describes how likely it is that the validated address is correct. Confidence code values range from 0 to 100, with a zero confidence level indicating no confidence and 100 indicating a very high level of confidence that the match results are correct. Confidence codes are calculated based on an algorithm that takes into account the match results for individual output fields. The output fields involved in this calculation include:
- Country
- City
- StateProvince
- PostalCode
- StreetName
- HouseNumber
- LeadingDirectional
- TrailingDirectional
- StreetSuffix
- ApartmentNumber
Each field has its own Weight in the algorithm. Additionally, for each field the match result could be labeled as Success, Failure, or Changed. ("Changed" refers to cases where the contents of the field have been corrected in order to get a match.) The match result—Success, Failure, or Changed—determines what the Factor is for that field. Thus, the calculation for the confidence code is a product of Weight by Factor as follows:
Confidence = (Weight * Factor) for City
+ (Weight * Factor) for Country
+ (Weight * Factor) for State
+ (Weight * Factor) for PostalCode
+ (Weight * Factor) for StreetName
+ (Weight * Factor) for HouseNumber
+ (Weight * Factor) for Directionals
+ (Weight * Factor) for Street Suffix
+ (Weight * Factor) for ApartmentNumber