LinkSolv 8.3 Help Pages and User Guide

How to Select Match Fields

 
Guidelines
The match specifications evaluate the candidate pairs from all passes in a single, consistent way by comparing data from Table A (Crash) with data from Table B (EMS).  They should include all available fields that can be made to fit the statistical model.  Match specifications remain the same for each pass.  It is OK to experiment with different specifications to find the best.  Fields from A and B must be paired in only one-way in your join and match specifications.  To pair a field multiple ways, copy the field multiple times with different names when you standardize.
 
Field from A:  Any field from standardized Table A (Crash).
Field from B:  Equivalent field from standardized Table B (EMS).
Compare Method:  Any available comparison method suitable for the type of data in the fields.
Low, Hi Limits:  The range of differences to be counted as agreements when field values are compared.
Agree, Disagree X:  The statistical model assumes exact comparisons and independent comparison outcomes when calculating match weights and match probabilities.  If you use comparisons with tolerances or fields with dependent outcomes then you must reduce agree and disagree weights appropriately as described in detail on page 79.  Otherwise you will increase the number of false positive and false negative matches.  You make adjustments by entering decimal fractions less than 1.  For example, entering 0.80 in the Agree X column means that only 80% of the calculated agree weight for that field is used to determine match probabilities.
Authored with help of Dr.Explain