I am using decision tree with a mix of continous and discrete variables to predict a continous variable.
Is there a procedure/function available to:
1. retrieve the score value generated by the lift chart>scatter plot
2. to mimic the GUI functionality under Mining Acccuracy Chart>Input Selection>Specify a different data set. For example the procedure would take in the Model Name, DataSet/View with the testing data and return the Lift Chart score for the testing
In summary, I am trying to call the 2 steps above programmatically instead of using the GUI..
i.e., of selecting a different dataset for a particular model to retrieve the lift chart score.
Using the scores, I intend to monitor the accuracy of the model continously over a period of time in the future to see if the data structure/rules have changed and model needs a retrain/rebuild.
View Complete Post