“…The algorithms evaluated to build the classifier are the C4.5 decision tree and a neural network with the multilayer perceptron (MLP) architecture. Both algorithms have been widely and successfully used in diabetes clinical domain with multiple purposes like retinal disease diagnosis (Bourouis, Feham, Hossain, & Zhang, 2014), impaired glucose metabolism or type 2 diabetes mellitus identification (Hische, Luis-Dominguez, & Pfeiffer, 2010;Mohlig, Floter, & Spranger, 2006;Shankaracharya, Mallick, & Shukla, 2012;Upadhyaya, Farahmand, & Baker-Demaray, 2013), or to identify significant factors influencing diabetes control (Huang, McCullagh, Black, & Harper, 2007). However, they have not been used for BG level classification in mealtime intervals.…”