Logo

Stanford Center for Biomedical Informatics Research Colloquia - Shared screen with speaker view
Pat Langley
21:07
What about bagging? That requires no weights.
jonathan chen
22:41
Post questions into the Q&A box and we'll relay to our speaker at the breaks
Pat Langley
24:29
I retract my question. The speaker doesn’t seem to be talking about ensembles after all.
Samuel Nuakoh
29:33
How do you begin exploration in an emergency situation [phase 3 + 1] on a global scale? At your first shot, just how many data sets are you required to acknowledge, accept and analyze? What range of numbers makes is scientifically and academicallyT good(C) to begin analysis.
yadevi
50:36
how we do decide the rsquared to input in the ronde. can we assume or from previous literature
Walid Al-Soneidar
56:30
Does the functional form of predictors have any effect on sample size requirement?
jonathan chen
58:31
Post questions to Q&A
Mary Eji
01:02:35
I'm not really clear about your response to the use of domain knowledge in sample size- would the analysis still be considered data-driven if domain knowledge is applied?
Mary Eji
01:05:42
What's your view about data augmentation ?
Mary Eji
01:12:24
Is an analysis of purely synthetic data a publishable project- in the medical domain particularly?
Mary Eji
01:14:05
Any threshold for imputation?
Minh Nguyen
01:18:30
thank you for that perspective!