Match specification - incorrect classification

Infosphere's Quality Product

Moderators: chulett, rschirm

Post Reply
vijaydasari
Participant
Posts: 29
Joined: Sun Jul 22, 2007 3:25 pm

Match specification - incorrect classification

Post by vijaydasari »

I am working on consolidating customer data exist in two different databases.

I setup match specification in below way.

block commnds are

MatchfirstnameNYSIIS_USNAME
MatchPrimaryname_USNAME
STATE_CD
ZIP CODE

match command

Matchfirstname_USNAME (Mprob 0.9 , Uprob 0.1 , Param1 850)
MatchPrimaryname_USNAME (Mprob 0.9 , Uprob 0.1 , Param1 850)
ADDRESS1 (Mprob 0.8 , Uprob 0.1 , Param1 700)
ZIP CODE (Mprob 0.6 , Uprob 0.1 , Param1 500)

Comparision type for all the columns is UNCERT

when I do TestALLPasses , below data combination showing as mater and duplicates record

firstname | Last Name | ZIP | STATE | address1 | weight
DAVID | Robin | 19382 | FL | 501 W Street Rd | 62.33
DAVID | Robin | 19382 | FL | 1119 S New St | 23.68

As per our requirement when there is no match in address , we need to consider as a different record . but now quality stage generating 1 master and duplicate record . Please help me how can i fix this?
Vijaya K Dasari
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Post by ray.wurlod »

Set your Match Cutoff above 23.68. Clearly that's not enough weight to be a match given your blocking fields and match rules.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
vijaydasari
Participant
Posts: 29
Joined: Sun Jul 22, 2007 3:25 pm

Post by vijaydasari »

In match designer window thre is cut off values section , do I need to set Match and Clerical section to 23.68?

is it advisable to add address1 column also into block section?

Please let me know
Vijaya K Dasari
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Post by ray.wurlod »

Sounds like you need to invest in some training, or at least in some time reading the manual. Match and clerical cutoffs serve different purposes. Please research that and post your findings back here.

Yes, I do know the answer, but my philosophy is to help you to help yourself. If I simply give you the answer you won't learn as effectively.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
vijaydasari
Participant
Posts: 29
Joined: Sun Jul 22, 2007 3:25 pm

Post by vijaydasari »

Hi Ray,

here are my findings . please let me know if i am wrong.

match cutoff is used identify match records . record pairs with compositer weight
more than cutoff are reliable match.


clerical cutoff is used identify posible match records . record pairs with compositer weight
more than clerical cutoff are possible match.

record pairs whose composite weight between match cutoff and clerical cutoff are clerical.
Vijaya K Dasari
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Post by ray.wurlod »

Almost but not quite. Records with composite weight below clerical cutoff are non-matches (residuals), at least for that pass. Otherwise good. Clerical cutoff must be set <= match cutoff.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
Post Reply