Knowledge uncertainty management in remote healthcare based on mutual information

Abstract

Providing primary healthcare for the people of rural India is a major challenge, even for common health issues such as cold, diarrhoea, etc. Different types of uncertainty often present in the health data and in remote areas scarcity of doctors and skilled manpower make the situation bad to worse. The paper aims to manage knowledge uncertainty using rough set theory (RST) and information theory by inducting patients from non-empty boundary to positive region based on the mutual information of the patients belong to the regions. Patients of the positive region are certainly diagnosed by the health workers and used as training samples to predict the new patients with certainty. To measure the performance of the proposed method, various measures are used. The results are verified using the ground truth of the experts.