Can Match cutoff be in negative?
I am trying to set it below 0 in Unduplicated Independent Stage in QS but match designer is not letting me enter anything other than 0 or positive value.
Match Cutoff Negative
-
- Premium Member
- Posts: 43
- Joined: Wed Feb 08, 2012 8:12 pm
- Location: United States
-
- Participant
- Posts: 54607
- Joined: Wed Oct 23, 2002 10:52 pm
- Location: Sydney, Australia
- Contact:
That is correct behaviour. Anything with an aggregate weight below zero is necessarily a non-match, since the sum of magnitudes of disagreement weights is greater than the sum of magnitudes of agreement weights.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
-
- Participant
- Posts: 527
- Joined: Thu Apr 19, 2007 1:25 am
- Location: Melbourne
Hey BuddingDev,
Because of the way it works out the base agreement and disagreement weights, abs(AW) > abs(DW).
So in order to get a negative overall score (without using weight overrides), you need to have a lot more non-matching fields than matching fields.
How confident would you be in a "match" where only 3 fields out of 10 matched and 7 were not? And I don't mean where there was missing data on 1 or both sides, but where both sides had data and it was different.
Because of the way it works out the base agreement and disagreement weights, abs(AW) > abs(DW).
So in order to get a negative overall score (without using weight overrides), you need to have a lot more non-matching fields than matching fields.
How confident would you be in a "match" where only 3 fields out of 10 matched and 7 were not? And I don't mean where there was missing data on 1 or both sides, but where both sides had data and it was different.
-
- Premium Member
- Posts: 43
- Joined: Wed Feb 08, 2012 8:12 pm
- Location: United States
-
- Participant
- Posts: 527
- Joined: Thu Apr 19, 2007 1:25 am
- Location: Melbourne
If you are using heavy weight overrides yet still feel that a score < 0 deserves a match, then it might be worth looking to see if there's another way get the result you're looking for.BuddingDev wrote:As a matter of fact, I was using weight overrides. So I was using heavy agreement or disagreement weight to fulfill the requirement.
-
- Participant
- Posts: 527
- Joined: Thu Apr 19, 2007 1:25 am
- Location: Melbourne
If you are using heavy weight overrides yet still feel that a score < 0 deserves a match, then it might be worth looking to see if there's another way get the result you're looking for.BuddingDev wrote:As a matter of fact, I was using weight overrides. So I was using heavy agreement or disagreement weight to fulfill the requirement.
-
- Premium Member
- Posts: 43
- Joined: Wed Feb 08, 2012 8:12 pm
- Location: United States
In an attempt to discard a record which shares the same info in different match sets in Unduplicated Matching, I have worked around the match cut off limit. I am receiving a weight of -99.99 for one of my master record for a particular sets which in reality has a lower weight than my given cut off.
Did any one has any experience with this kind of scenario?
Thanking in advance.
Did any one has any experience with this kind of scenario?
Thanking in advance.
This can happen when the master was determined in a previous pass, and then the overrides in a subsequent pass caused the -99 weight. It will still be output as a master because it did match on one of the passes.
Maybe if you posted more detail we could help you with what you are really trying to accomplish? Maybe the english description of the match you are trying to accomplish, and an example of what you are seeing that you do not want. A summary of the blocking/matching/overrides would probably help as well...
Maybe if you posted more detail we could help you with what you are really trying to accomplish? Maybe the english description of the match you are trying to accomplish, and an example of what you are seeing that you do not want. A summary of the blocking/matching/overrides would probably help as well...
Regards,
Robert
Robert
-
- Premium Member
- Posts: 43
- Joined: Wed Feb 08, 2012 8:12 pm
- Location: United States