Machine learning - longitudinal models (predict disease conversion)?

mtl-brainhack-school-2019 / rwickens-sMRI-PET

PET pipeline automation and structural MRI machine learning exploration

2 stars 0 forks source link

Machine learning - longitudinal models (predict disease conversion)? #7

Open rwickens opened 5 years ago

rwickens commented 5 years ago

I'm curious about statistical models that would predict the time of onset of disease conversion based on clinical measurements. I've seen the terms "survival analysis" and "cox proportional hazards" used. Does anyone have experience using such models? @PeerHerholz @atrophiedbrain

PeerHerholz commented 5 years ago

I more familiar with "classic" models, hence I could give you some hints there. However, as you mentioned, there are models which are tailored to specific applications, taking into account certain properties of the data (time series, feature importance, possible outcomes, priors, etc.). Wrt that I can't help you much. I would be happy to help exploring the model at hand once you found one you would like to try.