Bayesian networks for blood donor prediction

Abstract

Blood centers are responsible for managing blood stocks so that they satisfy at a considerable level, in addition to guaranteeing a quality standard with the blood collected. Both factors are possible if there is a control of regular donors. Thus, this article proposes a computational model that predicts regular blood donors, whose methodology joins the results of association measures to determine the most likely characteristics of a donor, with the Naive Bayes algorithm. The proposed model presented results superior to 68% accuracy and 73% precision in predicting a regular blood donor.