Abstract:
Malaria remains one of the major threats to public health and economic development, accounting for deaths of millions of people annually. Diagnosis of malaria fever has been inaccurate in many cases owing to factors such as decrease of medical work force, imbalance in doctor to patient ratio, human error and diagnostic kit error. A diagnostic model can be used to overcome the limitation of inaccurate diagnosis of malaria fever. Although, different researchers have worked on diagnosing malaria using different techniques but with a small dataset, which hinders the performance level of the model. Hence, the need for developing a diagnosis and treatment system that would assist in diagnosing malaria fever accurately. The aim of this research is to use Neural Network to develop a model for diagnosing malaria fever. In this research, a Multilayer Perceptron Neural Network (MLPNN) was adopted to formulate the model for diagnosing malaria fever. The experimental study of the proposed system was conducted using medical records of patients diagnosed of malaria in Mother and Child Hospital, Akure, Ondo State. The Malaria Diagnostic System was implemented using Visual Studio Code Editor Integrated Development Environment and XAMPP server. The input feature represents the symptoms of patients’ and the output was classified into four categories based on level of severity (Low, Mild, High and Severe). The performance of the model was evaluated using standard metrics such as Accuracy, Precision and Recall. The overall accuracy of the diagnostic system developed is 98.5%, which shows that using Neural Network with a large dataset helps in improving the performance of the system. The system can be adopted by health care practitioners especially those in rural areas where there are no sufficient doctors.
