Detecting periods in which a time series model fails to predict the observed volatility (Q1424645)
From MaRDI portal
 | This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use this page instead for the normal view: Detecting periods in which a time series model fails to predict the observed volatility |
scientific article; zbMATH DE number 2058969
| Language | Label | Description | Also known as |
|---|---|---|---|
| English | Detecting periods in which a time series model fails to predict the observed volatility |
scientific article; zbMATH DE number 2058969 |
Statements
Detecting periods in which a time series model fails to predict the observed volatility (English)
0 references
16 March 2004
0 references
Cumulative sums of squares (CUSUMSQ) and Bos-Ding-Fetherston (BDF) algorithms [\textit{T. Bos, D. Ding} and \textit{T. A. Fetherston}, Pacific-Basin Finance J. 6, 295--306 (1998)] for detection of multiple change-points in variances of independent variables are described. Results of simulation experiments to compare their performance are presented. The author proposes to apply CUSUMSQ to normal pseudo-residuals to check the fit of financial time-series models. As an example, the daily returns on Deutshe Bank shares (1974-1998) are considered.
0 references
cumulative sum of squares
0 references
CUSUM
0 references
CUSUMSQ
0 references
Bos-Ding-Fetherston algorithm
0 references
change-point detection
0 references
pseudo-residuals
0 references
0 references
0 references
