Creating an Easy ML experiment

This topic describes how to create an Easy ML experiment.

  1. Select Machine Learning > Easy ML.
  2. Click Create new.
  3. Under Details, specify a name for the experiment.
  4. Optionally, specify a description.
  5. Click Save.
  6. Upload your training dataset from the local system. The supported format is .csv.
  7. If the dataset contains a header row, leave the First row represents header check box selected. Clear the check box if the dataset does not contain a header row.
  8. Click Upload the data.
  9. Specify the problem type and select the Label column.
  10. Click Train Model.
  11. Select one of the trained models.
  12. Click Test the model.
  13. Upload the testing dataset from the local system. The supported format is .csv.
  14. Click Test model predictions.
  15. Click Save.
    The Generate Quest button is active after the Easy ML experiment steps have been completed successfully and the status of the experiment is Ready.
  16. Click Generate Quest to automatically generate a quest from your Easy ML experiment.
    The experiment status changes to Generating while the dataset and the quest are being created in the Datasets and Quests sections. The status of the experiment is displayed as Generated once the dataset and quest are successfully created.

    The dataset has a Data Loaded status. The quest has a Draft status.

  17. Optionally, click the Selected Model tab to preview the trained model you have selected as best in the model selection.
  18. Optionally, click the Discarded Models tab to preview the trained models you have discarded in the model selection step.