Sciences pures et appliquées
Creation of an AI-based diagnose app/algorithm with MATLAB for child patients (age 30 and under) with Type 1 and Type 2 Diabetes.
Benjamin De La Barrera
The reason why I wanted to do my project on this topic was because of my love for artificial intelligence, as well as my interest for creating and developing things with mathematical language (coding, algorithms). Choosing to do on medicine was a choice I found back in my third semester when I read an article about an AI-based diagnosing machine for dystonia, which used to be a disease which was very hard to diagnose. But, with the help of AI, the doctors were able to have a success rate of 98.8%. This helped me realized that I wanted to do something similar as my final project. As years passed by, diabetes has become a topic that is more and more open, which led to many AI implementations, but there are always many places that don’t have the technology, which leads to many misdiagnosed cases. By the end of my project, I was able to complete the app to the most I can, but I do want to work on it more to give more functions and to make it more useable altogether. To conclude, the app works very well, and the AI that watches over the user’s answers works very well with the app. The app’s accuracy was tested with random clinic cases from the diabetes clinic of Sainte-Justine Hospital, which had a result of 92.8% accurate, with 13 patients out of 14 being properly diagnosed.