Forgot to mention that we built two models because we have seen that patients with morbidities have very much different symptoms than healthier children. It is curious that in the case of children with morbidities our model has a prediction error of 5% while the other model has an 11% prediction error. These results aren't 100% reliable since the data set, in our opinion, is not rich enough.
Log in or sign up for Devpost to join the conversation.