gStream: Graph-Based Sequential Change-Point Detection for Streaming Data

Uses an approach based on k-nearest neighbor information to sequentially detect change-points. Offers analytic approximations for false discovery control given user-specified average run length. Can be applied to any type of data (high-dimensional, non-Euclidean, etc.) as long as a reasonable similarity measure is available. See references (1) Chen, H. (2019) Sequential change-point detection based on nearest neighbors. The Annals of Statistics, 47(3):1381-1407. (2) Chu, L. and Chen, H. (2018) Sequential change-point detection for high-dimensional and non-Euclidean data <arXiv:1810.05973>.

Package details

AuthorHao Chen and Lynna Chu
MaintainerHao Chen <[email protected]>
LicenseGPL (>= 2)
Package repositoryView on CRAN
Installation Install the latest version of this package by entering the following in R:

Try the gStream package in your browser

Any scripts or data that you put into this service are public.

gStream documentation built on May 6, 2019, 1:02 a.m.