Hi James / everyone, Thanks for the reply, I've been going through the wiki pages (which means that you can expect more questions in the future :-) )
May I ask if anyone has a nice set of test data that they use to try out this module. It's quite easy to write up some mock data, as I've been doing, but I guess that a well planned set of test data might do wonders... On Mon, Sep 26, 2011 at 8:06 PM, James Egg <[email protected]> wrote: > Hi,**** > > ** ** > > I can try to answer your questions about the matching output, though I’m > not sure why the groups file in the first example is just the one line. The > second example seems reasonable, though. The higher score for the pair 3,1 > is more likely a match than 3,2. The algorithm compares the fields marked > as “include,” as in “include this column in scoring.” In the second case, > the fields used for both pairs agreed on “jim,” but one pair has “A” > compared with “A” while the other pair had “A” compared with “B.” The > upward rating of 2 agreeing fields pushed the score of the pairs 3,1 above > the pair that had just one agreement and one disagreement.**** > > ** ** > > Have you looked at the wiki page for the module? It is at > https://wiki.openmrs.org/display/docs/Patient+Matching+Module**** > > ** ** > > There is information about the Swing GUI and some explanation of terms. > Between the OpenMRS module and the Swing GUI, the term “must match” means it > is a blocking column, and “should match” is included in scoring.**** > > ** ** > > James**** > > ** ** > > *From:* [email protected] [mailto:[email protected]] *On > Behalf Of *Suranga Kasthurirathne > *Sent:* Monday, September 26, 2011 4:52 AM > *To:* [email protected] > *Subject:* [OPENMRS-IMPLEMENTERS] Newbie questions on patient Matching > module**** > > ** ** > > > **** > > Hi all,**** > > ** ** > > I've been looking at the patient matching module standalone, and trying to > understand the matching concept.**** > > ** ** > > Assuming I have table A as follows,**** > > 0|1|2|3 (Column IDs)**** > > ---------**** > > 1|max|m|A**** > > 2|jack|m|A**** > > 3|jim|m|A**** > > ** ** > > and table B-**** > > 0|1|2|3 (Column IDs)**** > > ---------**** > > 1|jim|m|A**** > > 2|jim|m|B**** > > ** ** > > If I block on Column 1, and include column 2, then I would expect it to > identify that the third record in table A is similar to records 1 and 2 of > table B.**** > > After running the linkage process, I get the following results,**** > > ** ** > > the _groups.txt file says 0|0.0|3|2|jim|jim|m|m|A|B**** > > ** ** > > while the _new_match file says**** > > ** ** > > 0.0|3|1|jim|jim|m|m|A|A**** > > 0.0|3|2|jim|jim|m|m|A|B (i'm assuming that the double (0.0) in both > cols mean that both these records match evenly ?)**** > > ** ** > > - any idea why the _groups.txt file fails to identify the relationship > between 3|1 ?**** > > ** ** > > Instead if I were to run the same test blocking on Column 1, and including > columns 2 and 3,**** > > I would expect it to identify that the third record in table A is similar > to the first record of table B.**** > > ** ** > > the _groups.txt file says 0|3.1699250014423126|3|1|jim|jim|m|m|A|A**** > > ** ** > > while the _new_match file says**** > > ** ** > > 3.1699250014423126|3|1|jim|jim|m|m|A|A**** > > -3.1699250014423126|3|2|jim|jim|m|m|A|B**** > > ** ** > > - Can someone please help explain the data values given here > ? I'm assuming that this means that the first record (3|1) has a high > chance > of matching while the second record (3|2) has really low chance of doing > so > ?**** > > ** ** > > ** ** > > -- > Best Regards,**** > > ** ** > > Suranga**** > > ** ** > ------------------------------ > > Click here to > unsubscribe<[email protected]?body=SIGNOFF%20openmrs-implement-l>from > OpenMRS Implementers' mailing list > **** > ------------------------------ > Click here to > unsubscribe<[email protected]?body=SIGNOFF%20openmrs-implement-l>from > OpenMRS Implementers' mailing list -- Best Regards, Suranga _________________________________________ To unsubscribe from OpenMRS Implementers' mailing list, send an e-mail to [email protected] with "SIGNOFF openmrs-implement-l" in the body (not the subject) of your e-mail. [mailto:[email protected]?body=SIGNOFF%20openmrs-implement-l]
