[SOLVED] Rolling Up Data

maccten

Hi,

I have data that looks something like below

id attr1 attr2, attr3
a  0         1       0
a  1         0       0
a  0         0       0
a  0         0       1

I m using K-means to cluster this data but i want to cluster the IDs together so for example a,b,c would be clustered together because their attributes are the same
At the moment this isnt working as k-means is treating each row as a record as opposed to each id as a record

To get around this i was hoping that i could collapse four records (a single ID) onto one line

So for example

id, attr1, attr2, attr3
a     1       1        1

Is this possible within Rapidminer?

Thanks for your help

maccten

Hi,

The trick to fixing this problem was to use the aggregate operator and set the aggregation property to MAX around the label i wanted to roll up to in this case the ID

Thanks