Perhaps I am overlooking a solution, but is there a KNIME-based node for UMAP? After searching the community, it appears that examples have hopped into Python to make it work. That wouldn’t work for my use case, so I wanted to double-check if there is a native extension I can leverage.
Hi @Brock_Tibert -
Not that I am aware of. I did a quick check in our ticketing system and didn’t see any mention of it there either.
Maybe you could add a bit of detail about your use case and what you’d like to do, and then I could move this topic over to the Feedback & Ideas category?
I am (mostly) using KNIME to teach a variety of analytical concepts to MBAs. For a module on unsupervised ML, and more specifically, Dimensionality Reduction, I was looking to compare the output of PCA to other approaches. Below is a more formal summary:
As a user of KNIME, I need the ability to apply the UMAP algorithm for dimensionality reduction to my data. Similar to the dedicated node(s) for PCA, I need to be able to include a node in my workflow to perform this dimensionality reduction task without the need to install extensions or lean into programming languages (e.g. Python).
Acceptance Criteria:
- A UMAP node exists natively within KNIME.
- The algorithm is applied to the data
- The user has the option to drop the existing features and only include the newly created dimensions.
1 Like
Thanks for this. I’ve moved the topic over to Feedback & Ideas , which means the community can now vote on it. So, if this is something you’d like to see added, click the big yellow vote button at the top of the thread!
2 Likes