model’s performance on the train portion of the dataset

HAFEDH
HAFEDH Registered Posts: 10 ✭✭✭

Hi everyone,

I’m trying to figure out where I can find my model’s performance on the train portion of the dataset in Dataiku.

I can easily access the test set metrics in the model interface, but I don’t see where Dataiku shows the scores computed on the training set (AUC, F1, etc.).
Are these metrics available directly in the model results?

Thanks in advance for your help!

Answers

  • Alexandru
    Alexandru Dataiker, Dataiku DSS Core Designer, Dataiku DSS ML Practitioner, Dataiku DSS Adv Designer, Registered Posts: 1,356 Dataiker

    Hi,
    The metrics you see on Indeed are only on the test dataset.
    To compute these on the training dataset, you can use an evaluation recipe.



    1) If you did not use explicit train/test input, can either export the training dataset from here:

    attachment?token=eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJpZCI6MTE2MDc0Mjk4NSwiZG9tYWluIjoiZGF0YWlrdS5mcmVzaGRlc2suY29tIiwiYWNjb3VudF9pZCI6MTEyOTc5fQ.o43ePcklF4F-q03nOBX7MIlSF2ltT9uC5k7myuh8_rE

    2) Deploy the Model to the flow

    attachment?token=eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJpZCI6MTE2MDc0Mjk4NiwiZG9tYWluIjoiZGF0YWlrdS5mcmVzaGRlc2suY29tIiwiYWNjb3VudF9pZCI6MTEyOTc5fQ.1XVlwhHQmiRPdcT9MwF2K0zgGn0IlBEcyqEAYJIInzs

    3) Create an evaluation recipe on the training dataset.
    https://doc.dataiku.com/dss/latest/machine-learning/causal-prediction/evaluation.html

Setup Info
    Tags
      Help me…