I am trying to follow this step: Make predictions on an external dataset
I don't see RUN EXTERNAL getting populated after importing data with actuals values
Solved! Go to Solution.
My pleasure. I'm glad that it worked
It worked. Thank you.
They have to be equivalent values. You have Yes, no in your training, but 0,1 in your external data. Just change 0 to No, and 1 to Yes and it should work
Another thing. When you go to ROC Curve and you click on Data Selection do you see Add external test data ? If so, click on it before uploading your dataset.
I am using cloud based datarobot. target name in the external data is the same. tried csv file this time but no luck. Please look details below on how my target variable is in each data and what the values are: Please let me know if values should be 1,0 or Yes,nO In target variable in external_data.
train_data test_data external_data
Target variable: Churn Target variable: Churn Target variable: Churn
Values: No, Yes Values: left blank as we Values: 1,0
want model to predict
Are you using on-prem, or cloud base DataRobot? If on-prem, what version are you using?
If cloud, make sure the target name in the external dataset is the same. Check that excel didn't add extra characters, like white spaces that make the column name different. Also, try uploading a CSV file and see if that resolves the issue.
Hi, Thank you for your prompt response. I am following instructions from DataRobot docs.
Thank you for your prompt response.
My issue is: After uploading data, i am supposed to see options - Run External test, Compute Predictions. But i only see compute predictions.
my test_data_with_actuals has label values along with other features.
Thank you for your prompt response. My issue is: After uploading data, i am supposed to see options - Run External test, Compute Predictions. But i only see compute predictions.
my test_data_with_actuals has label values along with other features.
Here is how to import and evaluate external data:
1. Go to the model on which you want to evaluate your external data
2. Upload the external data
3. After the dataset is uploaded, you will see 2 options: Run external test, and Compute predictions, click on Run external test
4a. Now, if you want to see the global metric, go to Menu, and under the Leaderboard option, you will find Show external test column
By clicking on Show external test column, you should see a new column on your leader board
4b. If you check ROC Curve or Lift Chart, under Data Selection dropdown menu, under the EXTERNAL TEST title, you should see your dataset name
Please also check this
Is this what you are looking for?
Hey @sri1,
Which tutorial are you following?
If you want to make predictions on an external data set you can go
Here in the models tab Model->Predict->make Predictions:
And here in the deployments tab Deployment->Predictions->make predictions:
HTH,
Ira