Implement Clustering K-Medoids for new Data

[hitorijanai]

Hello guys,

I get some problems about Clustering using K-Medoids, already can create clustering with K-Medoids, but I confused how to implement that Cluster using new data.

Thanks in Advance,

Regards,

K-Medoid_Sample.knwf (35.5 KB)

[tommy]

Hey @hitorijanai,

you’re idea is to assign the defined cluster (on historical data) to new data?
This can be done by using the Joiner-Node and assign the “Cluster”-Column (see workflow attached).
In your example data you should consider the case where the combination in new data has no assigned cluster (no historical row).
K-Medoid_Sample.knwf (24.7 KB)

So you see, that with the clustering itself you already got the most part.
Hope this helps,
Greetz, Tommy