USING ARTIFICIAL NEURAL NETWORKS FOR PREDICTING NEW EPILEPSY CASES AT KWEKWE GENERAL HOSPITAL IN ZIMBABWE

Authors

  • Dr. Smartson P. Nyoni
  • Mr. Thabani Nyoni

Keywords:

case, volumes, predict, Zimbabwe

Abstract

Epilepsy is responsible for an enormous amount of “untold” suffering around the globe. Fortunately, effective and cost efficient treatment is available for the management of this neurological disorder. Unfortunately, in developing countries such as Zimbabwe, up to 90% of the people who have this condition, and sometimes even more, are excluded from care and consequently remain in the shadow of this treatment gap. In Zimbabwe, we have all reasons to consider epilepsy a healthcare priority. This study employed monthly time series data on epilepsy cases recorded and managed at Kwekwe General Hospital (KGH) from Janaury 2010 to December 2019, in order to predict epilepsy case-loads over the period January 2020 to December 2021. The popular ANN (12, 12, 1) model was applied. Residual analysis of this model indicated that it was really stable and thus suitable for predicting epilepsy case volumes at KGH over the out-of-sample period. The results of the study indicate that epilepsy cases will generally decline over the out-of-sample period. Amongst other recommendations, the study suggested the need for timely treatment of epileptic patients in order to save life.

Downloads

Published

2021-01-23

Issue

Section

Articles