Information about Generate ID attributes

nav

Hello,

I want to know how work the special attribute ID. In particular when I add this attribute to an exampleSet with the generate id attribute, is it the same ID added by the k-means clustering operator? I ask this because my problem is the follow: i made a clustering with k-means using a subset of features, but after clustering I want to explore the examples in each cluster also for the features not used for clustering. So what i done is adding an id to the original dataset and then join the clusteredSet with the original dataset by the ID, but I want to be sure that the two ID are the same in the generated one and in the clusteredSet. Thank you to all.

MariusHelf

You should add the id to the dataset (using Generate Id) before removing the features and doing the clustering. In that case, the k-Means operator won't generate an own id attribute, but use the one already present in the dataset. That way you can be sure that the clustered example set and the original dataset have consistent ids, because it's basically the *same* id attribute.

Cheers,
Marius