CART Algorithmus (Adult Data Set)

Hello everyone,

I’m still a beginner and need help implementing the CART algorithm in KNIME.
I’m looking for an Decision Tree Viewer that displays the decision tree graphically. Unfortunately the normal Decision Tree Viewer is not compatible with my Weka node.

Could someone take a look at my workflow and correct it if necessary?

Datensatz_Praesentation_final.knwf (31.4 KB)

Used Dataset: http://archive.ics.uci.edu/ml/datasets/adult

Many thanks in advance!

Hi Alex,
Welcome to KNIME!
I’m not sure whether there is a tree-viewer for WEKA-DT-models (If anyone else knows - would be glad to get this :slightly_smiling_face:). The output here is a rule-set (available via the DT-node). Two remarks:

  1. Data is csv, file-reader works well. I modified your workflow to inject the column names
  2. Added a column filter to “play” with the used attributes.

Results are quite OK, above 85% (stratified sampling) and ROC AUC above 87%

Greets from Steinheim / Germany

Mike
adult-weka.knwf (37.6 KB)

3 Likes

Thanks for your answer and help.

Regards,
Alex

This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.