Diagnosis of Malaria using Machine Learning Models

Malaria is a disease caused by the protozoan parasite plasmodium vivax, and spreads through the bites of infected mosquitoes. It causes close to 19,000 deaths annually in India alone, a testament to the truly devastating impact this disease can have if not treated properly. We recognize that an early diagnosis is paramount with regards to the treatment of any disease, especially one as potentially deadly as malaria. 

This work aims to facilitate the diagnosis of malaria by utilizing Machine Learning models to classify patients as infected or otherwise. The data we use to train our models consists of cell image data, more specifically, images of blood cells collected from patients. Using our predictive models, we attempt to accurately classify blood cells on the grounds of whether they are parasitized or uninfected. 

We also compare the performance metrics of the models we employ and seek to find the one that performs optimally and provides the most accurate predictions in the process. Since the data is not skewed, Classification Accuracy is a suitable metric for measurement of the performance of various models and will be used in our implementation. 

Keywords: Machine Learning, Malaria, Classification.