Explaining machine learning and how it can be used to help doctors

For a project in the doctoral course Using Maths and CS to do Social Good at Uppsala University, our goal was to make grade 7-9 students more interested in mathematics. To that end, we have created a video with an accompanying interactive blog post. In the video, we explain the essence of how machine learning works and how it can be used to help doctors discover heart attacks from patient ECGs. There is also a Swedish version of the video.

We tailor our explanation of machine learning to grade 7-9 students by basing it on the concept of straight lines (linear functions). We start by showing how a straight line can be fitted to some collected data and then used to make predictions, for example to predict ice cream sales from the outside temperature. Illustrative ice cream example

We then explain what ECGs are and how they are used by doctors when a patient comes to a hospital and complains about chest pain, trying to determine if the patient has a heart attack or not. Chest pain, ECG, heart attack?

Finally, we generalize our initial ice cream sales vs. outside temperature example to explain how machine learning can be used also to predict heart attack risk from a patient ECG, and how this can be used to help doctors discover patients who have heart attacks. Prediction of heart attack risk

Project group:
Daniel Gedon, Erik Hallström & Fredrik K. Gustafsson

comments powered by Disqus