Ensemble binary segmentation for irregularly spaced data with change-points

Abstract We propose a new technique for consistent estimation of the number and locations change-points in structure an irregularly spaced time series. The core segmentation procedure is ensemble binary method (EBS), which large multiple change-point detection tasks using are applied on sub-samples data differing lengths, then results combined to create overall answer. do not restrict total series can have, therefore, our proposed works well when spacings between short. Our main statistic time-varying autoregressive conditional duration model we apply transformation process order decorrelate it. To examine performance EBS provide simulation study various types scenarios. A proof consistency also provided. methodology implemented R package , available download from CRAN.