Malaria Disease Prediction Based on Machine Learning

Malaria detection is a stressful job for most doctors and it requires experiences and expertise. The machine learing (ML) method can be used to releave this issue. This paper try to find suitable model to help detect malaria with accuracy. The used dataset was released by National Institute of Health in USA and contained a total number of 27,560 red blood cell (RBC) images with equivalent instances of parasitized and uninfected RBCs images. A single hidden layer feedforward neural networks methodology namely Extreme Learning Machine (ELM) model was applied to classify and predict whether a patient has been affected by malaria or not. ELM has been compared with other machine learning techniques like SVM, KNN, CART, RF, CNN, VGG16, RESNET, and DENSENET, and it has outperformed all the other with 99% of accuracy, 28seconds cost time, 0.0095 Misclassification Error, and 98% precision which showed the effectiveness of ELM in the application of malaria cell detection scenario and it can also be referred by other researchers in the related field.