A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is
Sixth International Conference of Information Fusion, 2003. Proceedings of the
In the conventional formulation of the changepoint detection problem, there is a sequence of observations whose distribution changes at some unknown point in time, and the goal is to detect this change as quickly as possible, subject to false alarm constraints. It is known that in the case where the observations are independent and identically distributed (iid) and the change point is modeled as a geometrically distributed random variable, the Shiryaev detection procedure minimizes the expecteddoi:10.1109/icif.2003.177315 fatcat:7usk5ycnofhftjtpgvklerc2zi