>>12491440>How long does it take to make a machine learning model that takes in data from hospitals in the US and then tells me if a person is at risk for a heart attack?This is little too general, but you could make a model in like an hour, assuming you have the data. Will it be good? Meh.
But more generally:
1. What is "data from hospitals"? Are we talking people-risk factors, EKGs, something else? Is is a constant stream of data (ala every day new records are added) or are we taking, say, a large set of patient data from the past 10 years?
Do you have heart attack data from these people? The biggest thing about ML is how good the data is. Often times the data is collected shittily in a biased manner, dumped onto companies/contractors and they go "here do something with it." If you don't have the right data, its all bunk and useless.
I'm gonna assume you have a bunch of parameters/features and incident of heart attack in a big-ass data table. It would take a bit of time to clean it up, decide on a risk-factor model, and then train/test it. What, a week? I could do it in a day if I had nothing else on my plate, but like. That assumes the magical data table is already magically perfectly there.
Basically, the ML model is the easy part and can be run in like a day/hours, its the data clean-up/understanding what the fuck I'm looking at/figuring out if your shitty data actually makes for good predictions or not
