Artificial intelligence predicts heart attacks more accurately than the doctor

Cardiovascular diseases annually claim the lives of about 20 million people. While heart attacks are difficult to anticipate, despite the availability of modern diagnostic tools. Experts at Nottingham University (England) has developed smoooches a computer system that predicts the likelihood of a heart attack or stroke more accurately than does the doctor.

The recommendations of the American society of cardiology and the American heart Association (ACC/AHA), taking into account eight factors, provide forecast with an accuracy of 72.8 percent. The system created at the University of Nottingham, shows a figure of 74.5-76.4 per cent. While it is 1.6% rarely produces false signals.

The creators of the algorithm used for its training data from 295 000 378 256 patients available to them. The remaining records were used for verification. In short, improving the accuracy of prediction to the possibility of salvation 355 people from this group by using drug therapy or changes in diet. Note that artificial intelligence took into account factors not considered in the recommendations of the ACC/AHA.

Source: Science


(Visited 18 times, 1 visits today)
No tags for this post.