Scaglione, Marco (2021) Two P-spline applications to portfolio selection problems. [Tesi di dottorato]
Preview |
Text
Scaglione_Marco_33.pdf Download (33MB) | Preview |
| Item Type: | Tesi di dottorato |
|---|---|
| Resource language: | English |
| Title: | Two P-spline applications to portfolio selection problems |
| Creators: | Creators Email Scaglione, Marco marco.scaglione@unina.it |
| Date: | 19 July 2021 |
| Number of Pages: | 177 |
| Institution: | Università degli Studi di Napoli Federico II |
| Department: | Scienze Economiche e Statistiche |
| Dottorato: | Economia |
| Ciclo di dottorato: | 33 |
| Coordinatore del Corso di dottorato: | nome email Pagano, Marco marco.pagano@unina.it |
| Tutor: | nome email Aria, Massimo UNSPECIFIED |
| Date: | 19 July 2021 |
| Number of Pages: | 177 |
| Keywords: | P-spline; Functional Data Analysis; Portfolio Selection |
| Settori scientifico-disciplinari del MIUR: | Area 13 - Scienze economiche e statistiche > SECS-S/01 - Statistica |
| Date Deposited: | 19 Jul 2021 18:53 |
| Last Modified: | 07 Jun 2023 11:16 |
| URI: | http://www.fedoa.unina.it/id/eprint/13558 |
Collection description
In this manuscript I provide two applications of an established smoothing technique, the Penalized Splines smoother, to solve the roughness issue in the econometric analysis of the portfolio selection problem. In the first chapter I face the problem of statistical hedge ratio estimation through quantile regression, here the goal of the P-spline application is to avoid model specification and to smooth the quantile regression objective function. In the second chapter the smoother is applied in a time series filtering framework in order to achieve smoothness in Principal Component Analysis and provide smoother principal component for regression analysis. This allows robust feature selection from principal component loadings in order to perform portfolio selection with index tracking purpose in an high dimensional context.
Downloads
Downloads per month over past year
Actions (login required)
 |
View Item |
