The current AI system has shown great promise in diagnosing disease, analyzing medical images, and predicting health outcomes, even better than doctors in terms of surgical suturing and diagnosing infant autism. But now, with new advances in AI medical applications, researchers at the University of Nottingham have created a system to scan patients' routine medical data to predict their risk of heart disease or stroke over the next 10 years. It is actually a very difficult task to predict such cardiovascular disease. In a recent paper, researchers say that about half of heart attacks and strokes occur in people who are not labeled as "at risk." The current standard method for assessing the risk of a patient's illness depends primarily on guidelines developed by the Society of Cardiac Association. The current standard focuses on the risk of hypertension, cholesterol, age, smoking and diabetes. Researcher Stephen Weng and his colleagues tested several different machine learning tools based on the medical records of 378,256 patients in the UK. These medical records record the patients and their health status from 2005 to 2015, including medical conditions, prescription drugs, hospital visits, and test results. The researchers sent 75% of their medical records to their machine learning model to identify the characteristics of patients who experienced a heart attack or stroke within 10 years. Model tests were then conducted on the other 25% of the records to determine how accurately they predicted heart attacks and strokes. If 1.0 is used to indicate 100% accuracy, the traditional forecasting standard score is 0.728. The machine learning model results from 0.745 to 0.764, and the best score comes from the neural network machine learning model: the neural network model successfully predicted 4,998 cases in 7404 actual cases, 355 more than the traditional method. Using this technology to predict can help doctors take appropriate precautions, such as prescribing drugs to patients at risk of developing a disease to lower cholesterol. So how does the AI ​​tool help doctors in the actual diagnosis? Stephen Weng said that their algorithm can mark patients with risk of onset after reviewing and analyzing the entire patient list to alert doctors. This process can occur either when the patient is sitting in front of the doctor for routine examinations or when the patient is not present. Stephen Weng pointed out that the main advantage of the platform is predictive accuracy: although similar clinical decision support software already exists, unlike the software, they developed systems that use AI pattern recognition to provide more accurate predictions of results. Hand massager Hand massager Shenzhen Jie Zhong Lian Investment Co., Ltd. , https://www.meizons.com