2
$\begingroup$

I have recently used the heatmap widget in Orange 3. All the documentation says is "Clustering (clusters data by similarity)". Is this using hierarchical or k-means or some other type of clustering?

On that note, is there a way to look at the code being run by all the widgets to see whats going on under the hood? It would be nice if after you finish the workflow you would get a file with the script run to perform the analysis.

$\endgroup$
0
$\begingroup$

It appears the widget uses hierarchical clustering. I guess the metric is Euclidean distance by default and there doesn't seem to be a way to specify another one (except by using Distances widget and connecting it into the Distance Map widget).

I don't think it is possible to export the widget's workflow as pure code, but you can look at what the widget does in the source code (seems pretty low-level, though). What you can do, however, is select subsets of data (can be saved with Save Data widget) for further analysis if that's of any help.

Orange Heatmap widget screenshot

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.