The sample pair file

Within the sample pair .xls file are rows of member data grouped in pairs.

Most of the columns are self-explanatory, shown in the format Attribute Name | Field Label. The specific attributes shown in the file depend on what was selected when you ran the Generate Threshold Analysis Pairs job. Other columns of importance are:

• Same? - indicates whether the members in each pair are the same member. One or both members of each pair should be marked with Yes, No, or Maybe. This column is not displayed in the InfoSphere® MDM Pair Manager.
• Pair number - numbers each pair of members sequentially.
• Score - the comparison score assigned by InfoSphere MDM. The higher the number, the greater the likelihood the two members in the pair represent the same individual.
• Original score - the first comparison score assigned by the operational server. When members are rescored following an algorithm update, their score often changes. The original score, however, remains static.
• memrecno - the member record number in the MDM database assigned to the member.
• Source - the source system where the member data originated.

If you prefer to use a spreadsheet application to evaluate the sample pairs, see the instructions in Generate Threshold Analysis Pairs.

If the algorithm is changed after running Generate Threshold Analysis Pairs, you can rescore the sample pairs by using the Virtual MDM > Rescore members in sample pairs file(s) menu. See Rescoring members in sample pair files.