Consider a large number of detectors each generating a data stream. The task is to detect online, distribution changes in a small fraction of the data streams.
We propose optimal algorithms that minimize the detection delay subject to a given average run length constraint.
We also compare these algorithms against previous approaches that include the use of mixture likelihood ratios and sum of CUSUMs.
1. Optimal sequential detection in multi-stream data. Working paper. Chan, H. (2016).