Finding Periodic Discrete Events in Noisy Streams

Abhirup Ghosh, Christopher Lucas, Rik Sarkar
2017 Proceedings of the 2017 ACM on Conference on Information and Knowledge Management - CIKM '17  
Periodic phenomena are ubiquitous, but detecting and predicting periodic events can be di cult in noisy environments. We describe a model of periodic events that covers both idealized and realistic scenarios characterized by multiple kinds of noise. e model incorporates false-positive events and the possibility that the underlying period and phase of the events change over time. We then describe a particle lter that can e ciently and accurately estimate the parameters of the process generating
more » ... eriodic events intermingled with independent noise events. e system has a small memory footprint, and, unlike alternative methods, its computational complexity is constant in the number of events that have been observed. As a result, it can be applied in low-resource se ings that require real-time performance over long periods of time. In experiments on real and simulated data we nd that it outperforms existing methods in accuracy and can track changes in periodicity and other characteristics in dynamic event streams.
doi:10.1145/3132847.3132981 dblp:conf/cikm/GhoshLS17 fatcat:q7y7io3eajgehc7fcto7misk5q