SAP Message RSDME304 - No valid decision tree formed


Diagnosis
The data used for training does not contain sufficient or the requiredinformation to form a decision tree. This could be due the followingreasons:
  • Model parameters like Min. Leaf Cases or Leaf Node Accuracy

    • are set to very high values
    • Specified model fields in the model have no strong relation with the
      • predictive model field values
      • The data has too many missing values

        • System Response
        Procedure
        Change model parameter values and perform training.
        If there are some fields in the data that could have an effect on theprediction model field value, include these fields in the modeldefinition (if they are not already included). Perform training again.
        If you have used pruning, deactivate it and re-train the model. Theresulting tree may be prone to over-fitting and hence may not beaccurate.
        If you have tried all of the above and you still do not obtain a validdecision tree, this implies that the data you are using is not fit fordecision tree training.
        Procedure for System Administration