Blood glucose prediction for diabetes therapy using a recurrent artificial neural network

Abstract

Expert short-term management of diabetes through good
\nglycaemic control, is necessary to delay or even prevent
\nserious degenerative complications developing in the long
\nterm, due to consistently high blood glucose levels
\n(BGLs). Good glycaemic control may be achieved by
\npredicting a future BGL based on past BGLs and past and
\nanticipated diet, exercise schedule and insulin regime (the
\nlatter for insulin dependent diabetics). This predicted
\nBGL may then be used in a computerised management
\nsystem to achieve short-term normoglycaemia. This paper
\ninvestigates the use of a recurrent artificial neural network
\nfor predicting BGL, and presents preliminary results for
\ntwo insulin dependent diabetic females.