Abstract

This study proposes the development of a data analytics framework for diagnosis of prevalent illness among university students. A high-level model methodology with a Cross Industry Standard Process for Data Mining (CRISP-DM) steps was adopted. The findings showed that "Plasmodiasis" also known as "Malaria" has the highest occurrence, followed by "body pain" and then "Flu", while the forty-five (45) other illnesses have less or insignificant amount of occurrence. For Malaria to be prevalent there is need to deal with it in ways that would drastically reduce it occurrence or rate of hospitalisation, and also the distraction from lectures. Engagement of the model in diagnosis of prevalent illness will allow evidence-based awareness on the prevalent illness, personalised treatment; reduce human-prone errors using sample data from the Federal University Lokoja Health Centre. Furthermore, another finding showed that the Gradient Boosting Classifier had 100% accuracy, 100% precision and 100% recall as compared to other six algorithms.