problem on loop attribute subsets
I need to conduct experiment using loop attribute subsets, but I do not need all combination of the subsets.
The experiments just need to add one attribute for every iteration.
For example: we have 10 attributes (attr_1, attr_2, ..., attr_10).
We only need to run a model on attr_1, {attr_1, attr_2}, {attr_1, ... attr_3}, ..., {attr_1, attr_2, ..., attr_10}
Is it possible to conduct such experiments in RapidMiner?
The experiments just need to add one attribute for every iteration.
For example: we have 10 attributes (attr_1, attr_2, ..., attr_10).
We only need to run a model on attr_1, {attr_1, attr_2}, {attr_1, ... attr_3}, ..., {attr_1, attr_2, ..., attr_10}
Is it possible to conduct such experiments in RapidMiner?
Tagged:
0
Answers
Here's the outline of an answer, it runs in the reverse order to your requests, because it starts with all attributes, and trashes the last one at each iteration. I've stuck a break in the loop to make the point, so just keep pressing on ;D