K-CUSUM: Cluster Detection Mechanism in EDMON

Abstract

The main goal of the EDMON (Electronic Disease Monitoring Network) project is to detect the spread of contagious diseases at the earliest possible moment, and potentially before people know that they have been infected. The results shall be visualized on real-time maps as well as presented in digital communication. In this paper, a hybrid of K-nearness Neighbor (KNN) and cumulative sum (CUSUM), known as K-CUSUM, were explored and implemented with a prototype approach. The KNN algorithm, which was implemented in the K- CUSUM, recorded 99.52% accuracy when it was tested with simulated dataset containing geolocation coordinates among other features and SckitLearn KNN algorithm achieved an accuracy of 93.81% when it was tested with the same dataset. After injection of spikes of known outbreaks in the simulated data, the CUSUM module was totally specific and sensitive by correctly identifying all outbreaks and non-outbreak clusters. Suitable methods for obtaining a balance point of anonymizing geolocation attributes towards obscuring the privacy and confidentiality of diabetes subjects’ trajectories while maintaining the data requirements for public good, in terms of disease surveillance, remains a challenge.