Hierarchical clustering a long dataset of distances

KNIME Analytics Platform
#1

Hi,

I'm a Knime noob so apologies if my question is a really basic one:

I have a long dataset of "distances" between pairs of entities, effectively an unpivoted distance matrix:

Entity1, Entity2, Distance
A, B, 0.2
A, C, 0.4
B, C, 0.5

I'd like to hierarchically cluster this data.

I had expected the "Hierarchical Clustering (DistMatrix)" node to do this but the "Distance matrix column" is disabled (and I don't have anything connected to input port1).

I'm sure this must be an easy one.
Please can anyone suggest a solution?

Thanks,

Will

#2

Hi Will,

for using the Hierarchical Clustering you need a distance matrix column and this you can for example generate  with the Distance Matrix Calculate.

We also have a blog post about this: https://www.knime.org/blog/distance-measures-in-knime

 

Best, Iris