東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Autoregressive Conditional GB2 Model...

Fan, Ning.

FindBook

Google Book

Amazon

博客來

Autoregressive Conditional GB2 Models with Applications to Financial Time Series.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Autoregressive Conditional GB2 Models with Applications to Financial Time Series./
作者:	Fan, Ning.
出版者:	Ann Arbor : ProQuest Dissertations & Theses, : 2020,
面頁冊數:	105 p.
附註:	Source: Dissertations Abstracts International, Volume: 82-07, Section: B.
Contained By:	Dissertations Abstracts International82-07B.
標題:	Statistics. -
電子資源:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28263800
ISBN:	9798557043045

Autoregressive Conditional GB2 Models with Applications to Financial Time Series.
Fan, Ning.

Autoregressive Conditional GB2 Models with Applications to Financial Time Series. - Ann Arbor : ProQuest Dissertations & Theses, 2020 - 105 p.

Source: Dissertations Abstracts International, Volume: 82-07, Section: B.

Thesis (Ph.D.)--The University of Wisconsin - Madison, 2020.

This item must not be sold to any third party vendors.

In order to model the time varying behavior of maxima in financial time series, we introduce a novel dynamic generalized beta distribution of second kind (GB2) framework. The proposed autoregressive conditional GB2 (AcGB2) model will be fitted to the high skewed and long tailed financial time series data. The time dependence among the maxima is characterized in the parameter dynamics of the GB2 distribution. One merit property of GB2 distributions is that many distributions including extreme value distributions can be approximated by GB2 distributions with different parameter values. Incorporating the dynamics on the parameters results in the time series of maxima possessing distribution dynamics, e.g., observations can be at extreme levels and at less extreme levels (intermediate extremes). As a result, the newly proposed modeling framework has greater flexibility than extreme value distributions and can better be fitted to real data. The proposed thesis work proves the existence of stationary and ergodic solution of the new time series model under mild conditions of the parameters. Statistical estimation method (conditional maximum likelihood estimation) will be developed. The consistency, the asymptotic normality, and the uniqueness of estimators will be derived. The thesis uses simulation examples to demonstrate the model properties and the efficiency of statistical inference method. Real data analysis will be applied to two data sets, the maxima of negative logarithm returns from 30 stocks in Dow Johns Industrial Average (DJIA), and the maxima of negative logarithm returns from 101 stocks in S&P 100. The thesis also compares the AcGB2 method inference with the existing method AcF in terms of their performance in real applications.

ISBN: 9798557043045Subjects--Topical Terms:

517247
Statistics.
Subjects--Index Terms:

Dynamic modeling

Autoregressive Conditional GB2 Models with Applications to Financial Time Series.
LDR:02869nmm a2200337 4500 001 2278454
005 20210628075031.5
008 220723s2020 ||||||||||||||||| ||eng d
020 $a 9798557043045
035 $a (MiAaPQ)AAI28263800
035 $a AAI28263800
040 $a MiAaPQ $c MiAaPQ
100 1 $a Fan, Ning. $3 3385756
245 1 0 $a Autoregressive Conditional GB2 Models with Applications to Financial Time Series.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2020
300 $a 105 p.
500 $a Source: Dissertations Abstracts International, Volume: 82-07, Section: B.
500 $a Advisor: Zhengjun, Zhang.
502 $a Thesis (Ph.D.)--The University of Wisconsin - Madison, 2020.
506 $a This item must not be sold to any third party vendors.
520 $a In order to model the time varying behavior of maxima in financial time series, we introduce a novel dynamic generalized beta distribution of second kind (GB2) framework. The proposed autoregressive conditional GB2 (AcGB2) model will be fitted to the high skewed and long tailed financial time series data. The time dependence among the maxima is characterized in the parameter dynamics of the GB2 distribution. One merit property of GB2 distributions is that many distributions including extreme value distributions can be approximated by GB2 distributions with different parameter values. Incorporating the dynamics on the parameters results in the time series of maxima possessing distribution dynamics, e.g., observations can be at extreme levels and at less extreme levels (intermediate extremes). As a result, the newly proposed modeling framework has greater flexibility than extreme value distributions and can better be fitted to real data. The proposed thesis work proves the existence of stationary and ergodic solution of the new time series model under mild conditions of the parameters. Statistical estimation method (conditional maximum likelihood estimation) will be developed. The consistency, the asymptotic normality, and the uniqueness of estimators will be derived. The thesis uses simulation examples to demonstrate the model properties and the efficiency of statistical inference method. Real data analysis will be applied to two data sets, the maxima of negative logarithm returns from 30 stocks in Dow Johns Industrial Average (DJIA), and the maxima of negative logarithm returns from 101 stocks in S&P 100. The thesis also compares the AcGB2 method inference with the existing method AcF in terms of their performance in real applications.
590 $a School code: 0262.
650 4 $a Statistics. $3 517247
650 4 $a Finance. $3 542899
653 $a Dynamic modeling
653 $a Extreme values
653 $a Financial risk management
690 $a 0463
690 $a 0508
710 2 $a The University of Wisconsin - Madison. $b Statistics. $3 2101047
773 0 $t Dissertations Abstracts International $g 82-07B.
790 $a 0262
791 $a Ph.D.
792 $a 2020
793 $a English
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28263800