Angaben zur Quelle [Bearbeiten]
| Autor | M.-O. Boldi |
| Titel | Mixture models for multivariate extremes. |
| Ort | Lausanne |
| Jahr | 2004 |
| Anmerkung | Dissertation |
| URL | http://biblion.epfl.ch/EPFL/theses/2004/3098/EPFL_TH3098.pdf |
Literaturverz. |
nein |
| Fußnoten | nein |
| Fragmente | 4 |
Fragmente der Quelle:
| [1.] Rh/Fragment 011 11 - Diskussion Zuletzt bearbeitet: 2012-07-31 15:39:46 KnallErbse | Boldi 2004, Fragment, Gesichtet, KeinPlagiat, Rh, SMWFragment, Schutzlevel |
|
|
|
| Untersuchte Arbeit: Seite: 11, Zeilen: 27 |
Quelle: Boldi 2004 Seite(n): 16, Zeilen: 9 |
|---|---|
| The multivariate theory is naturally more recent. | The multivariate theory is naturally more recent. |
Sehr kurz, daher keine Bewertung. Inhaltlich einschlägige Dissertation. |
|
| [2.] Rh/Fragment 016 20 - Diskussion Zuletzt bearbeitet: 2012-08-15 22:17:23 Hindemith | Boldi 2004, Fragment, Gesichtet, Rh, SMWFragment, Schutzlevel sysop, Verschleierung |
|
|
|
| Untersuchte Arbeit: Seite: 16, Zeilen: 20-22 |
Quelle: Boldi 2004 Seite(n): 20, Zeilen: 23-25 |
|---|---|
| In practice, the estimation in (2.2.9) requires one to decompose the sample into blocks and take the blockwise maxima. A drawback with this approach is the loss of data. | In practice, the estimation of μ, σ and κ requires one to decompose the sample into blocks and take the blockwise maxima. A drawback with this approach is the loss of data. |
Ein Quellenverweis fehlt. |
|
| [3.] Rh/Fragment 017 17 - Diskussion Zuletzt bearbeitet: 2012-07-31 16:15:36 WiseWoman | Boldi 2004, Fragment, Gesichtet, KomplettPlagiat, Rh, SMWFragment, Schutzlevel sysop |
|
|
|
| Untersuchte Arbeit: Seite: 17, Zeilen: 15-18 |
Quelle: Boldi 2004 Seite(n): 22-23, Zeilen: S.22,23-25 - S.23,1-2 |
|---|---|
| The Peaks Over Threshold method can be represented as a semi-parametric model. The excesses above a high threshold u are distributed according to a GPD, while the empirical distribution function of F, or any other appropriate model, is used under the threshold u. This is the semi-parametric extremal model, see for example Coles and Tawn (1991). | [Seite 22]
The Peaks Over Threshold method can be represented as a semi-parametric model. The excesses above a high threshold u are distributed according to a generalized Pareto distribution while the empirical distribution function F, or any other appropriate model, is [Seite 23] used under u. This is the semi-parametric extremal model, see for example Coles & Tawn (1991). |
fast identisch ohne Quellenangabe; lediglich "generalized Pareto distribution" wurde durch "GPD" ersetzt. |
|
| [4.] Rh/Fragment 019 23 - Diskussion Zuletzt bearbeitet: 2012-08-10 09:23:57 Hindemith | Boldi 2004, Fragment, Gesichtet, Rh, SMWFragment, Schutzlevel sysop, Verschleierung |
|
|
|
| Untersuchte Arbeit: Seite: 19, Zeilen: 23-28 |
Quelle: Boldi 2004 Seite(n): 27, Zeilen: 2-7 |
|---|---|
| In the absence of this condition, the dependence structure is such that large value has a greater chance of being followed by another one. If the time between two consecutive such values is small relative to n, the passage to the limit will merge those two extremes onto the same time. Thus, the limit process is then not a Poisson process but a compound Poisson process: any occurrence can be multiple rather than single. The multiplicity is usually random and is called the cluster size distribution π(∙). | In the absence of condition D', the dependence structure is such that a large value has a greater chance of being followed by another one. If the time between two consecutive such values is small relative to n, the passage to the limit will merge those two extremes onto the same time. The limit process is then not a Poisson process but a compound Poisson process: any occurrence can be multiple rather than single. The multiplicity is usually random and is called the cluster size distribution. |
Ein Quellenverweis fehlt |
|