Home > Research > Publications & Outputs > Asymptotically consistent estimation of the num...

Research

Associated organisational unit

Links

View graph of relations

Asymptotically consistent estimation of the number of change points in highly dependent time series

Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSN › Conference contribution/Paper › peer-review

Published

Standard

Asymptotically consistent estimation of the number of change points in highly dependent time series. / Khaleghi, Azedeh; Ryabko, Daniil.
Proceedings of The 31st International Conference on Machine Learning. 2014. p. 539-547.

Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSN › Conference contribution/Paper › peer-review

Harvard

Khaleghi, A & Ryabko, D 2014, Asymptotically consistent estimation of the number of change points in highly dependent time series. in Proceedings of The 31st International Conference on Machine Learning. pp. 539-547. <http://jmlr.org/proceedings/papers/v32/khaleghi14.html>

APA

Khaleghi, A., & Ryabko, D. (2014). Asymptotically consistent estimation of the number of change points in highly dependent time series. In Proceedings of The 31st International Conference on Machine Learning (pp. 539-547) http://jmlr.org/proceedings/papers/v32/khaleghi14.html

Vancouver

Khaleghi A, Ryabko D. Asymptotically consistent estimation of the number of change points in highly dependent time series. In Proceedings of The 31st International Conference on Machine Learning. 2014. p. 539-547

Author

Khaleghi, Azedeh ; Ryabko, Daniil. / Asymptotically consistent estimation of the number of change points in highly dependent time series. Proceedings of The 31st International Conference on Machine Learning. 2014. pp. 539-547

Bibtex

@inproceedings{8bbac98f9e154c848877d8b3786d6424,
title = "Asymptotically consistent estimation of the number of change points in highly dependent time series",
abstract = "The problem of change point estimation is considered in a general framework where the data are generated by arbitrary unknown stationary ergodic process distributions. This means that the data may have long-range dependencies of an arbitrary form. In this context the consistent estimation of the number of change points is provably impossible. A formulation is proposed which overcomes this obstacle: it is possible to find the correct number of change points at the expense of introducing the additional constraint that the correct number of process distributions that generate the data is provided. This additional parameter has a natural interpretation in many real-world applications. It turns out that in this formulation change point estimation can be reduced to time series clustering. Based on this reduction, an algorithm is proposed that finds the number of change points and locates the changes. This algorithm is shown to be asymptotically consistent. The theoretical results are complemented with empirical evaluations.",
author = "Azedeh Khaleghi and Daniil Ryabko",
year = "2014",
month = jun,
language = "English",
pages = "539--547",
booktitle = "Proceedings of The 31st International Conference on Machine Learning",

}

RIS

TY - GEN

T1 - Asymptotically consistent estimation of the number of change points in highly dependent time series

AU - Khaleghi, Azedeh

AU - Ryabko, Daniil

PY - 2014/6

Y1 - 2014/6

N2 - The problem of change point estimation is considered in a general framework where the data are generated by arbitrary unknown stationary ergodic process distributions. This means that the data may have long-range dependencies of an arbitrary form. In this context the consistent estimation of the number of change points is provably impossible. A formulation is proposed which overcomes this obstacle: it is possible to find the correct number of change points at the expense of introducing the additional constraint that the correct number of process distributions that generate the data is provided. This additional parameter has a natural interpretation in many real-world applications. It turns out that in this formulation change point estimation can be reduced to time series clustering. Based on this reduction, an algorithm is proposed that finds the number of change points and locates the changes. This algorithm is shown to be asymptotically consistent. The theoretical results are complemented with empirical evaluations.

AB - The problem of change point estimation is considered in a general framework where the data are generated by arbitrary unknown stationary ergodic process distributions. This means that the data may have long-range dependencies of an arbitrary form. In this context the consistent estimation of the number of change points is provably impossible. A formulation is proposed which overcomes this obstacle: it is possible to find the correct number of change points at the expense of introducing the additional constraint that the correct number of process distributions that generate the data is provided. This additional parameter has a natural interpretation in many real-world applications. It turns out that in this formulation change point estimation can be reduced to time series clustering. Based on this reduction, an algorithm is proposed that finds the number of change points and locates the changes. This algorithm is shown to be asymptotically consistent. The theoretical results are complemented with empirical evaluations.

M3 - Conference contribution/Paper

SP - 539

EP - 547

BT - Proceedings of The 31st International Conference on Machine Learning

ER -