Publication

Abstract : Most of the advanced computer models that predict serious illnesses, like a brain stroke, are "black boxes." Although this information may give advance notice of potentially at-risk individuals, it does not divulge the basis of their predictions. That opaqueness is a huge barrier to entry because it makes doctors hesitant to trust and leverage such powerful tools. This study fills an important void in that it provides one of a kind AI solution that delivers high accuracy and is straightforward to interpret. We introduce a new model, the Learnable Rule-Based Neural Network (LRBNN). LRBNN designs the core logical rules based on prior medical knowledge regarding the age, blood pressure, and other clinical factors, thus be-forming an extremely intuitiveness model.