FYI - Academy: Machine Learning Basics error

MarkPundurs
MarkPundurs Dataiku DSS Core Designer, Dataiku DSS ML Practitioner, Registered Posts: 27 ✭✭✭✭

Dataiku Support has forwarded the following ticket to the Academy team:

Course module https://academy.dataiku.com/path/ml-practitioner/machine-learning-basics/546062 states that "Patient Number" is the most important feature in the model (at about 1:10 in the video). At a minimum, the feature needs to be renamed to better convey its meaning; if this feature is really just an ID as the current name suggests, then any predictive power is at best a proxy for some more meaningful variable (e.g., time since discharge, which would negatively correlate with probability of readmission). Suggesting that IDs should be used as features is a grave disservice to inexperienced modelers.

Answers

  • Sean
    Sean Dataiker, Alpha Tester, Dataiku DSS Core Designer, Dataiku DSS ML Practitioner, Dataiku DSS Adv Designer Posts: 168 Dataiker

    Hi @MarkPundurs
    , we've logged this as something to look into. Thanks for flagging it for us.

  • taraku
    taraku Dataiker, Alpha Tester, Dataiku DSS Core Designer, Dataiku DSS ML Practitioner, Registered Posts: 53 Dataiker

    @MarkPundurs
    thank you for reporting this issue. We are constantly striving for the highest level of quality in our courses. Soon we will release an updated version of this module where we have decided not to include "Patient Number" as a feature in training the model. Stay tuned for these updates and please continue to communicate any findings by using our feedback form!

Setup Info
    Tags
      Help me…