How can we deal with unknown member in cubes for data mining?

Hi, all here,

I currently have a question about unknown member in cubes (as we know, there is always an additional member called unknown in an OLAP dimension)for data mining. I mean is there any way for us to deal with unknown member in cubes when mining cubes?

Thanks a lot in advance for any guidance and help.

The unknown member maps to the missing state when you build a mining structure based on an OLAP cube and this is usually the behavior you want. Are you looking for a different behavior?|||

Hi, Raman, thanks a lot.

But my question is the unknown member in cubes is just added by the system automatically? Since the unknown member is actually not existing in the original data sources? Is there any way to delete the unknown member in cubes before data mining? Thanks a lot for guidance.

|||Yes the unknown member is added automatically during dimension processing. You can change the UnknownMember property of a dimension to None in BI Dev Studio by bringing up the dimension properties and expanding "Advanced" in the property pane. You'll need to reprocess the dimension and affected cubes.|||

Hi, Raman, thanks a lot. I have followed the steps you guided, but the unknown member is still shown in the cube content though?

1. set the property of unknown member to be none.

2. reprocess the cube.

But when I browse the cube, the unknown member is still shown in the cube though.

Thank you very much for your further advices.

|||The property change should work. Did you re-process the dimension as well? The property change may not be a "structural change" that forces dimension re-processing when the cube is reprocessed.|||Hi, Raman, thanks a lot.