Heart smart: a novel deep learning approach to improving heart disease diagnosis

Sofia Tomov, Tennessee, USA 13-15

My project uses artificial intelligence to improve the accuracy of heart disease diagnosis. Heart disease is the #1 killer worldwide, and research shows that doctors' basic screening tools can be inaccurate and slow. To address this problem, I coded HEARO: Heart Evaluation for Algorithmic Risk-reduction and Optimization. The HEARO software uses a novel combination of optimizations to analyze 13 common metrics of patient data and output the probability that an individual has heart disease. It achieves 99% accuracy, outperforming previous results, including those from Stanford. This can provide doctors with a potentially lifesaving tool to make a more accurate diagnosis.

Show more