Comments on: Entity Resolution Metrics http://identityresolutiondaily.com/654/entity-resolution-metrics/ All About Identity and Entity Resolution Fri, 10 Feb 2012 01:42:34 +0000 http://wordpress.org/?v=2.2 By: John Talburt http://identityresolutiondaily.com/654/entity-resolution-metrics/#comment-718 John Talburt Mon, 23 Nov 2009 13:28:05 +0000 http://identityresolutiondaily.com/654/entity-resolution-metrics/#comment-718 Jim, Thank you. That is an interesting question that I have not fully considered. It is clear that if E is order dependent then at most only one of the partitions that it produces can be correct. However it is not clear if an order independent process E would necessarily have a better chance of producing a "more correct" partition. My first thought is that you could construct examples that work both ways, but I will think about this further. -jrt- Jim,
Thank you. That is an interesting question that I have not fully considered. It is clear that if E is order dependent then at most only one of the partitions that it produces can be correct. However it is not clear if an order independent process E would necessarily have a better chance of producing a “more correct” partition. My first thought is that you could construct examples that work both ways, but I will think about this further.
-jrt-

]]> By: Jim Zaiss http://identityresolutiondaily.com/654/entity-resolution-metrics/#comment-715 Jim Zaiss Sat, 21 Nov 2009 08:01:17 +0000 http://identityresolutiondaily.com/654/entity-resolution-metrics/#comment-715 Suppose that, for a given S, there is a process E such that each possible ordering λi always results in the same partition P. Can we then conclude – ceteris paribus -- that P is more likely to be the _correct_ partition of S than another partition P’ based on another process E’ such that the λi chosen sometimes results in a different partition of S? Please excuse me if this is a naïve question, or if it describes a situation that rarely occurs in practice. I’m new to this particular topic, but find it fascinating. I’ve long been interested in personal identity issues, though mostly from a logical, metaphysical, or phenomenological perspective. Regards, Jim Zaiss AWARE Software, Inc. Suppose that, for a given S, there is a process E such that each possible ordering λi always results in the same partition P. Can we then conclude – ceteris paribus — that P is more likely to be the _correct_ partition of S than another partition P’ based on another process E’ such that the λi chosen sometimes results in a different partition of S?

Please excuse me if this is a naïve question, or if it describes a situation that rarely occurs in practice. I’m new to this particular topic, but find it fascinating. I’ve long been interested in personal identity issues, though mostly from a logical, metaphysical, or phenomenological perspective.

Regards,
Jim Zaiss
AWARE Software, Inc.

]]>