Nodes for finding Best K values in clustering and Node for Significant variables

Hi ,
Which node is useful to Select best K value for clustering in 4.6.5 version and which node is uesd to select significant variables for building linear regression.
Please help with this asap. thanks in advance .

For clustering try either Elbow Method or Silhouette method:

I’ll get back to you on regression.


feature selection for variable selection

1 Like

More than one way to kill the proverbial cat.

Try this workflow for ranking factors. There are more sophisticated methods invloving differential calculus, but either of these approaches should be adequate.

1 Like

currently i have updated to 4.7.3 version, I am not able to find the nodes in updated version. and if possible could someone share workflows for finding optimum clusters using elbow method in K_means and PCA.

Thanks in advance.

Look above. I provided links to workflows for elbow and silhouette methods.

Yeah i have gone through it but am not able to find those nodes and extensions in KNIME 4.7.3 version, please help to install extensions and find those nodes.

What are you missing? The workflows should include everything you need. Have you tried to run them? If you want to create your own workflow, you can copy what you need from the workflows I provided you.

Ohh is it, but don’t we need to install extensions if i want to use those nodes?
How can i get the nodes in Node repository.

Hey, Actually i am able to use the nodes directly by copying, as We need to populate the workflows for learners in order to provide them with the necessary tools and resources. Could you please provide guidance on how we can access those nodes from the node repository? Alternatively, if there is a specific source within KNIME where we can obtain the required nodes, please kindly mention it.

Thanks in advance.

What learners? Just curious. What are you preparing?
If you add the nodes to a KNIME workflow and export it so others can open it they will be automatically ask to install missing extensions

Silhouette Method:
Here are links to the node and component used in the two workflows:

Here are the revised workflows which employ the best cluster # to calcuate a final k-means:

Elbow method:
There are no “nodes” which calculate an adjusted R^2. I changed the metanode used in the workflow to a component.

A revised workflow which calculates the final k-means based on the best cluster # is here:


Thanks for sharing the nodes. as we are an ed-tech company and we are using this for teaching purpose so, can you please provide what are the extensions required to get the nodes that you had shared

Nothing else should be necessary. Try running the workflows I sent on your machine. They should run.

Can you share company name? As @rfeigel said, openening the flows should ask for installation automatically.

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.