USING ARTIFICIAL NEURAL NETWORKS FOR PREDICTING NEW SNAKE BITE CASES AT GWERU PROVINCIAL HOSPITAL IN ZIMBABWE

Abstract

Snakes are well known to human race, as they continue to be associated with a plethora of epics, myths, superstitions, folklore, and tales and so on. In Zimbabwe, just like in any part of the world, when a snake it seen, it is usually killed. Snakes are also considered as sacred and sometimes certain rituals are performed when certain snakes are found. Unfortunately, snake bites remain a source of considerable morbidity and mortality in many parts of the world, especially in developing countries such as Zimbabwe. The current study used monthly time series data on snake bite caseloads recorded and managed at Gweru Provincial Hospital (GPH) from Janaury 2010 to December 2019, to predict snake bite cases over the period January 2020 to December 2021. We applied the ANN (12, 12, 1) model. Residual analysis of the applied model indicates that the model is stable and thus suitable for forecasting snake bite case volumes at GPH over the out-of-sample period. The results of the study reveal that snake bite cases will generally be on a downwards trajectory at GPH over the out-of-sample period. The study, amongst other policy suggestions, encourages the GPH management team to always make sure that there is prompt administration of antisnake venom (ASV) to victims in order to save life from this preventable public health threat.