Artificial Intelligence Approach to Symptom-Based Disease Prediction and Department Routing

Abstract

In contemporary hospital appointment systems, patients often book appointments without having any idea about the nature of their ailment. Incorrect appointment bookings have the potential to negatively impact both the patient and the hospital in terms of time and cost. Additionally, patients are often unclear about which symptoms correspond to which diseases and which hospital departments provide relevant services for their conditions. In this study, by giving users the right to choose among 132 predefined different disease symptoms, the diseases the user has among 41 different diseases are predicted through an artificial intelligence model created with the Multinomial Naive Bayes algorithm. In the study, the user is presented with the top three disease predictions with the highest probability based on the symptoms selected in an interface developed with the React library. Additionally, the study aims to inform the user by providing information about the relevant hospital departments and explanations of the diseases.