Orange Forum • View topic - On using predicted values in plots

On using predicted values in plots

A place to ask questions about methods in Orange and how they are used and other general support.

On using predicted values in plots

Postby alokg » Mon Jan 13, 2014 5:32

Hi all,

I am using scatter plot widget on data passing through k-means clustering. The original data has class labels to be learned, and the k-means method assigns a cluster to each sample.

In the scatter plot, the point color can be linked to the cluster, but not class label. These are both discrete variables, so I am wondering why is this arbitrary limitation placed. After some searching around, I found that the class labels and clusters assigned can be swapped using "Select Attributes" widget, but I still do not follow the motivation for this. Can someone please explain?


Return to Questions & Support