Using Artificial Intelligence in Diagnostics of Meningitis

Abstract

Meningitis is a serious condition caused by inflammation of meninges which are protective layers of membranes for the brain and spinal cord. The aim of this study is to help diagnose meningitis using artificial intelligence. In this study we used a database consisting of data collected from 1000 subjects, which were grouped in two groups. One group consisted of 800 diseased subjects and the other group consisted of 200 healthy subjects. Parameters used in the diagnosis of meningitis were: temperature, CSF/blood glucose ratio, proteins, CSF leukocytes, glucose, lactates, erythrocyte sedimentation rate and CRP. Based on the training results, a neural network with 15 hidden layer neurons was chosen for diagnosis of meningitis. Artificial neural networks have a total accuracy of 96.69%, which means it can be successfully used for diagnosis of meningitis.