The AI ​​system can predict a stroke risk more accurately than a doctor

1. What happens if aliens are artificial intelligence?
2. Stephen Hawking revealed the AI ​​artificial intelligence solution that dominates people
3. Microsoft Research is creating an AI capable of writing code in seconds

According to current estimates, about 20 million people die each year from cardiovascular diseases . Fortunately, a team of researchers at Nottingham University, England, has developed a machine learning algorithm that can predict a patient's ability to have a heart attack or stroke with parity, even lice are more than today's real doctors.

The AI ​​system can predict a stroke risk more accurately than a doctor Picture 1

American College of Cardiology / American Heart Association ( American College of Cardiology / American Heart Association - ACC / AHA ) has successfully developed a series of guidelines that predict cardiovascular risk for patients based on 8 weaknesses. factors, including age, cholesterol levels and blood pressure. This AI system can help physicians predict an average of 72.8% of cases.

The AI ​​system can predict a stroke risk more accurately than a doctor Picture 2

Although this number is very positive, Stephen Weng and his colleagues at Nottingham University want to improve their accuracy even further. They built 4 machine learning algorithms for computers to acquire, learn the medical data of 378,256 patients across the UK.

Initially, the system used about 295,000 medical records to produce a predictive model. Then use the remaining medical records to test and improve the model. The machine learning algorithms developed by Stephen's team have yielded accurate estimates that surpass the guidelines of the American Heart Association - AHA, from 74.5 to 76.4%.

The AI ​​system can predict a stroke risk more accurately than a doctor Picture 3

Of the 83,000 tested cases, the AI ​​system could have saved an additional 355 human lives if used earlier. In addition, the interesting thing here is that computer algorithms identify a number of factors that contribute to increased cardiovascular risk that have not been mentioned in the AHA guidelines as patients with a history of Serious psychological illness or excessive use of oral corticosteroids, for example.

It is clear that the current computer technology science is not only in the hardware and software products with normal performance but also helps us exploit more unknown things about our own bodies.