BACK to VOLUME 31 NO.1

Kybernetika 31(1):17-29, 1995.

A Method of Detecting Changes in the Behaviour of Locally Stationary Sequences

Jiří Michálek


Abstract:

A method for the detection of abrupt changes in the course of a locally stationary sequence is proposed. The method is based on a suitable approximation of an observed sequence by autoregressive models that are compared by means of a similarity measure derived from the asymptotic $I$-divergence rate. The method is illustrated by several numerical results.


download full-text.pdf
download abstract.pdf

BIB TeX

@article{kyb:1995:1:17-29,

author = {Mich\'{a}lek, Ji\v{r}\'{\i}},

title = {A Method of Detecting Changes in the Behaviour of Locally Stationary Sequences},

journal = {Kybernetika},

volume = {31},

year = {1995},

number = {1},

pages = {17-29}

publisher = {{\'U}TIA, AV {\v C}R, Prague },

}


BACK to VOLUME 31 NO.1