Hi to everybody,
I am new in Knime, therefore, I need your valuable support 
I am clustering a time series and now I want to visualize my results. I was thinking about it this way where you have a multiple line plot of different profiles identified from the cluster. How can I implement this in Knime? Is there any combination of nodes that I can do it?
Hi Matthew, I don’t have much knowledge about time-series clustering, but I guess you can check the clustering algorithms that are available, such as the Knime natives ones:
or the Weka clustering nodes
Assuming that the above clustering algorithms are suitable, then you can employ the Line Plot node to compare the results.
Hope it helps
Time_Series_Clustering.knwf (90.0 KB)
Mau
Hi @Matthew313, complementing @mauuuuu5’s answer, you can assign a color to your prototype using an idea that was suggested on this thread how to apply colors to different columns by @AlexanderFillbrunn
Color Column.knwf (802.7 KB)
@mauuuuu5 thank you for your response, yes I am using DBSCAN in my case and I am trying to figure it out as time series is a little bit complex as you have the time component that should be always in the visualization even though you are not clustering it (as a variable).
Thank you @iperez for your response, it seems very helpful
I will try it out immediately
This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.
