Outbreak detection for temporal contact data

Abstract Epidemic spreading is a widely studied process due to its importance and possibly grave consequences for society. While the classical context of epidemic refers pathogens transmitted among humans or animals, it straightforward apply similar ideas spread information (e.g., rumor) computer viruses. This paper addresses question how optimally select nodes monitoring in network timestamped contact events between individuals. We consider three optimization objectives: detection likelihood, time until detection, population that affected by an outbreak. The approach we use based on simple greedy has been proposed seminal focusing water contamination. extend this work setting disease present application with two example networks: sexual contacts animal transports farms. procedure large set outbreak scenarios generate susceptible-infectious-recovered model. find heuristic methods high degree many compare well terms performance (greedily) optimal nodes. Furthermore, observe optimized past periods may not be future periods. However, seasonal effects help determining which period generalizes some period. Finally, demonstrate depends simulation settings. In general, if force simulator larger outbreaks, will improve, as outbreaks tend occur more connected part where top are typically located. A natural progression analyze representative can generated, taking into account realistic propagation models.