Variable Clustering Usage

KNIME Analytics Platform
1

Hello, KNIME Community,

I would like to perform variable clustering in KNIME.

Instead of clustering observations (rows), my goal is to cluster variables (columns).
For example, when working with a dataset containing around 100 columns, I would like to group similar variables together and select representative variables from each cluster in order to reduce the number of features.

Are there any KNIME nodes or recommended workflows that support this type of variable/feature clustering?

Any guidance or suggestions would be greatly appreciated.
Thank you in advance.

2

You can try a Correlation Filter or Feature Selection Loop or Principal Component Analysis. Any of these can remove redundant features. These are just general suggestions. Which might work best depends on your data and what you’re trying to do.

1 Like
3

Considering the functionalities provided by KNIME, possible approaches would be the Feature Selection Loop, Correlation Filter, or PCA.

1 Like
4

@JaeHwanChoi you can take a look the passage about “Data Preparation” in my article about KNIME and Machine Learning:

KNIME — Machine Learning and Artificial Intelligence — A Collection