Can "dirty" data help predict life-threatening events? #ai #datascience #healthai #datascientist

Indexed: 2026-05-22

[00:00:00]Even with empirical evidence that these types of models can work on on garbage data sets, I myself received pushback from hospital administrators, clinicians, um who were kind of unable to kind of accept that a high-stakes prediction, stroke, heart attack, could be made accurately from the hospital records. And part of that is because they know what goes into the hospital record data collection. It's it's dirty, it's messy, lots of things are coded incorrectly. There's lots of missing and part of that is just like people believe garbage in, garbage out without thinking about kind of why and what's behind that. Um