Extract components from Sepctral Decomposition in the PCA Node - Look up rowIds

Dear all,

I have just started using KNIME to analyse a larger dataset using the PCA Compute node. I am specifically interested in identifying the component with the highest loading in the respective eigenvectors (as supplied in the Spectral Decomposition table output).

What I have done so far is transpose this output, so I have one column for each eigenvector and the components as RowIDs. I am now looking for a way to detect e.g. the maximum loading for a given Eigenvector, but in terms of outputting the RowID (thus the component in my case) where this maximum is found. 

Is there a way to "look up" RowIDs by the value or property of the data contained in the cells?


Thank you for your help,






Hi Joachim,

I am not sure what the maximum loading is... 

But one solution would be to find the nearest neighbor in the original data set? 

This could be done with first the numerical distance node and than the similiarity search node.

I hope this helps.

Best regards, Iris