Dear committee,
Thank you for the event. I have some questions on the Event_time.
In the "Additional notes for Final Round - Test Phase", it is described that Event and Event_time will be removed from scoring dataset.
https://www.synapse.org/#!Synapse:syn27130803/wiki/620471
However, in the recent mail "Final Submission IMPORTANT" sent on 1/19, it is described that we should "use the Event_time information and submitted a classification model."
Could you please explain more on it?
Should we use Event_time as one of predicting targets to evaluate the performance of the model?
Namely, when building model, we should predict 2 targets "Event" and "Event_time, " but Event_time is not one of the input features and we need to use the model to estimated the heart failure risk (float 0~1), am I right?
It would be much appreciated if you could clarify, thank you very much!
Best regards,
Robin
Created by Chih-Han Huang Chih-Han Thanks for clarification! Dear Robin,
Please be aware that the challenge's goal is not to build a classification model. Instead, we are looking for models considering both event status and time to event information. Models that only focus on classifying Event information will not be accepted. We have noticed that some teams have not used the Event_time information, which is a requirement for any survival modelling when developing the model.
To clarify, we would like participants to develop models that predict the time-to-event for heart failure, not simply the Event itself (which is a classification problem, ).
Best wishes, ece