Table of Contents

Search

  1. Introduction
  2. Configuring Hub Console Tools
  3. Building the Data Model
  4. Configuring the Data Flow
  5. Executing Informatica MDM Hub Processes
  6. Configuring Application Access
  7. MDM Hub Properties
  8. Viewing Configuration Details
  9. Search with Solr
  10. Row-level Locking
  11. MDM Hub Logging
  12. Table Partitioning
  13. Collecting MDM Environment Information with the Product Usage Toolkit
  14. Glossary

Lightweight Matching

Lightweight Matching

If you want to improve match performance, configure lightweight matching. Lightweight matching generates extremely fast score estimates. The algorithm rejects candidates that contain obvious mismatches instead of passing them to full scoring. On a typical data set, the algorithm rejects greater than 99% of the candidates, which results in improved performance.
If you already use tight exact-match rules on columns that contain values like identification numbers or date of birth, lightweight matching might not provide significant performance improvement.
SSA-NAME3 rejects the candidates based on the lightweight matching score. It is possible for SSA-NAME3 to reject the candidates that might have matched high with the SSA-NAME3 scoring. You can mitigate this risk by carefully selecting the fields to which you apply lightweight matching and by using threshold tuning.
Use the LWM_FIELDS control to apply lightweight matching to the fields. Use the LWM_LIMIT control to set the reject and accept limits for the lightweight matching score.

Configuring Lightweight Matching

Configure lightweight matching by setting properties in the
cmxcleanse.properties
file:
For example, the following example enables lightweight matching on the specified fields.
cmx.server.match.lwm=true cmx.server.match.lwm_param=LWM_FIELDS=Organization_Name,50,Address_Part1,50 LWM_LIMIT=75,85 cmx.server.match.stats=false
The following property definitions describe how to use the lightweight matching properties together:
cmx.server.match.lwm
Optional. Must be added manually. Enables the lightweight matching feature, which uses SSA-NAME3. Default is
false
.
Use this parameter with the
cmx.server.match.lwm_param
and
cmx.server.match.stats
properties.
cmx.server.match.lwm_param
Optional. Must be added manually. Requires that the
cmx.server.match.lwm
property is set to
Y
or
ONLY
. Add the SSA-NAME3 controls in the following format:
LWM=Y LWM_FIELDS=<field1>,<weight1>[,...,<fieldn>,<weightn>] LWM_LIMIT=<Reject>[,<Accept>]
cmx.server.match.stats
Optional. Must be added manually. Requires that the
cmx.server.match.lwm
property is set to
Y
or
ONLY
.

SSA-NAME3 Controls

You can add the following SSA-NAME3 controls within the
cmx.server.match.lwm_param
property. Separate the controls with a space.
LWM=Y/N/ONLY
Enables or disables lightweight matching. Use the value
Y
to enable lightweight matching. Lightweight matching uses a fast score estimate to reject the obvious mismatches. The records that lightweight matching passes go to the full scoring for robust scoring and ranking. SSA-NAME3 returns the full score and the decision to the caller.
If you create system definition files by using the SDF Wizard, the lightweight matching is enabled by default.
Use the value
N
to disable lightweight matching. SSA-NAME3 matching performs full scoring on all the matching records.
Use the value
ONLY
to enable lightweight matching and disable full scoring. Lightweight matching returns the estimate as the final score to the caller.
LWM_FIELDS
Specifies the fields to which you want to apply lightweight matching and their weights. These values override the values that you have defined in the match purpose during the run time. Based on the lightweight matching scores, SSA-NAME3 rejects the obvious mismatches. If you do not set any value, SSA-NAME3 retrieves the fields from the match purpose and assigns equal weight to them.
The syntax of the LWM_FIELDS control is as follows:
LWM_FIELDS=<field1>,<weight1>[,...,<fieldn>,<weightn>]
where
field
is a valid field name that you have defined in the Purpose control, and
weight
is the relative significance of the specified field (0-100) when compared to the other fields.
For example,
LWM_FIELDS=Person_Name,5,Address_Part1,1
Lightweight matching is useful when you apply it to the fields that have low variations such as addresses. Lightweight matching is not efficient for the fields with high variations, where SSA-NAME3 handles the variations through Edit-list, and lightweight matching might incorrectly reject the records.
LWM_LIMIT
Specifies the accept and reject limits for the lightweight matching score. Based on the limits, SSA-NAME3 accepts or rejects the search results.
The syntax of the LWM_LIMIT control is as follows:
LWM_LIMIT=<Reject>[,<Accept>]
where
Reject
and
Accept
are the integer values ranging from 0 through 100.
For example,
LWM_LIMIT=50,90
If
LWM=N
, the
LWM_LIMIT
control has no effect.
If
LWM=Y
, SSA-NAME3 rejects the lightweight matching scores that are less than the reject limit. The accept limit has no effect, and you can omit it.
If
LWM=ONLY
, SSA-NAME3 rejects the lightweight matching scores that are less than the reject limit. It accepts the scores that are greater than the accept limit. It marks the scores of the records that are greater than or equal to the reject limit and less than the accept limit as undecided.
The default reject limit is 65, and the default accept limit is 90. If you have not set the accept limit and the reject limit is greater than 90, the accept limit is equal to the reject limit.

0 COMMENTS

We’d like to hear from you!