LinkSolv 8.3 Help Pages and User Guide
(c) 2012 Strategic Matching, Inc.

Match Passes Tab

Match Passes Tab
1

Where am I?

1. Where am I?
 
Match Passes -- Guidelines
The order in which passes are run does not affect merged results. You might want to run the pass with the fewest candidate pairs first to confirm that results look realistic. Incomplete passes are not included in Merge Passes. View Pass displays a datasheet view of the Match Pairs Pass N table.
2

Match Pass List

2. Match Pass List
 
Run a Pass. Select a pass to run. Click on Run Pass or click on the Run button for the selected pass.
 
Re-Run a Pass. Select a pass to re-run. Uncheck Complete. Click on Run Pass or click on the Run button.
 
Skip a Complete Pass. Select a complete pass to skip in Merge Passes. Check the Skip box.
3

Auto Run button

3. Auto Run button
 
Auto Run. If you have already completed all of the steps once then you can click on Auto Run to repeat Import Data and Standardize.
4

Run Pass button

4. Run Pass button
 
Run Pass. Select a pass to run. Click on Run Pass or click on the Run command button.
 
5

View Pass button

5. View Pass button
 
View a Complete Pass. Select a complete pass to view. Click on View Pass.
 
 
 
  1. Table Listing Candidate Pairs. One table is produced for each match pass. Candidate pairs with match weights greater than the cutoff weight are listed in a table and pairs below the cutoff are dropped. The table name is set to the Match Name with the Pass Name as a suffix, for example CrashHospital__Pass1. If Access is the Data Provider, the table is stored in a database with the same name as the table.
  2. Match Weight. Total match weight after considering all comparison outcomes. Weights are calculated using your prior estimates for all model parameters. Agree weights are averages for each match field, not value specific weights.
  3. Match Probability. Posterior probability that the listed pair is true match. Calculated from the match weight using Bayes' rule.
  4. UniqueID. Record identifier for a record in Table A in a match of Table A to Table B.
  5. UniqueID_B. Record identifier for a record in Table B in a match of Table A to Table B.
  6. RandomNbr. A random number between 0 and 1 used in Monte Carlo calculations.
  7. CrashCounty1__Exact. There is one column in the table for each match field. The column name is the field name from Table A with the Compare Method from the match specifications as a suffix.
Authored with help of Dr.Explain