8 problems
- 0 votes0 replies0 views
Optimal window-size tradeoff for private online community-change detection
In online change detection for dynamic community structures under a censored block model, let denote the network size and let the window size determine the amount of data used…
- 0 votes0 replies0 views
Suitability of the MTFA-based false-alarm constraint for independent non-stationary models
Let the mean time to false alarm (MTFA) be the reciprocal of the false-alarm rate (FAR) used to constrain a sequential change-detection procedure. Consider independent, non-station…
- 0 votes0 replies0 views
Periodic-threshold optimality conjecture for minimax quickest change detection
In a periodic independent, identically distributed model, consider quickest change detection under Lorden's minimax criterion, and allow stopping algorithms whose thresholds vary p…
- 0 votes0 replies0 views
Non-optimality of single-threshold policies for minimax change detection
Consider a non-Bayesian or minimax version of the change-detection problem studied in the paper, and let the false alarm rate vary over its possible values. A single-threshold poli…
- 0 votes0 replies0 views
Existence of a classical Shiryaev example with differing algorithm performance
The classical Shiryaev formulation compares the periodic-Shiryaev algorithm with the classical Shiryaev algorithm for detecting a change, using false-alarm and delay penalties. Exi…
- 0 votes0 replies1 view
Third-order asymptotic optimality of the Shiryaev–Roberts procedure
Third-order asymptotic optimality conjecture. For the Pollak min-max criterion, when is initialized with , the stopping rule is third-order asymptotically…
- 0 votes0 replies0 views
The inequality for the Shiryaev–Roberts optimality condition
Let be the function defined by the integral equation in the paper, let be the positive constant determined by … and let . The inequality conject…
- 0 votes0 replies0 views
Optimality of Low-Sum-CUSUM with corrupt sensors
Let sensors monitor for a change, of which at most may be corrupt. The Low-Sum-CUSUM procedure stops when the sum of the smallest local CUSUM statistics crosses a fix…