The Annals of Statistics
- Ann. Statist.
- Volume 43, Number 1 (2015), 139-176.
Graph-based change-point detection
We consider the testing and estimation of change-points—locations where the distribution abruptly changes—in a data sequence. A new approach, based on scan statistics utilizing graphs representing the similarity between observations, is proposed. The graph-based approach is nonparametric, and can be applied to any data set as long as an informative similarity measure on the sample space can be defined. Accurate analytic approximations to the significance of graph-based scan statistics for both the single change-point and the changed interval alternatives are provided. Simulations reveal that the new approach has better power than existing approaches when the dimension of the data is moderate to high. The new approach is illustrated on two applications: The determination of authorship of a classic novel, and the detection of change in a network over time.
Ann. Statist. Volume 43, Number 1 (2015), 139-176.
First available in Project Euclid: 18 November 2014
Permanent link to this document
Digital Object Identifier
Mathematical Reviews number (MathSciNet)
Zentralblatt MATH identifier
Primary: 62G32: Statistics of extreme values; tail inference
Chen, Hao; Zhang, Nancy. Graph-based change-point detection. Ann. Statist. 43 (2015), no. 1, 139--176. doi:10.1214/14-AOS1269. https://projecteuclid.org/euclid.aos/1416322039.
- Supplementary material: Supplement to “Graph-based change-point detection”. Supplement A: Proofs for lemmas, propositions and theorems. We provide the proofs to the lemmas, propositions and theorems. Supplement B: Skewness correction. We provide the details to the skewness correction we used. Supplement C: Checking analytic approximations to $p$-values. We provide more results in checking analytic approximations to $p$-values. Supplement D: Block permutation results. We provide more results on block permutation in analyzing the two real data examples.