Projects per year
Abstract
Understanding changes plays a crucial role in many fields of science, economy, technology and medicine. Whenever time series data, i.e., temporally ordered data, is described by means of a stationary stochastic model, it is of interest to verify the stationarity assumption on the basis of the data, and change point analysis provides mathematical tools for this purpose. A particularly important, and thus widely studied, problem is to detect changes in the mean at unknown time points. In this paper, we present the R package mosum, which implements elegant and mathematically welljustified procedures to themultiple mean change problem using moving sum (MOSUM) statistics.
Original language  English 

Journal  Journal of Statistical Software 
Publication status  Accepted/In press  8 Sep 2019 
Fingerprint Dive into the research topics of 'mosum: A package for moving sums in change point analysis'. Together they form a unique fingerprint.
Projects
 1 Finished

Changepoint detection for highdimensional time series with nonstationarities
25/06/16 → 24/12/17
Project: Research