Μεταπτυχιακές Εργασίες
Μόνιμο URI για αυτήν τη συλλογήhttps://beta-pyxida.aueb.gr/handle/123456789/63
Νέα
Αυτή είναι η συλλογή από το παλιό σύστημα με ID:cid:3
Περιηγούμαι
Πλοήγηση Μεταπτυχιακές Εργασίες ανά Συγγραφέα "Arvanitis, Stylianos"
Τώρα δείχνει 1 - 2 από 2
- Αποτελέσματα ανά σελίδα
- Επιλογές ταξινόμησης
Τεκμήριο Empirical assessment of Value-at-Risk methodologies on financial markets(2021) Μιχελάκης, Χαράλαμπος-Παναγιώτης; Athens University of Economics and Business, Department of International and European Economic Studies; Tzavalis, Elias; Arvanitis, Stylianos; Dendramis, YiannisWe study the prominent risk measures used in the field of finance for the purpose of forecasting extreme, but rare, adverse events; this is done in order to ensure capital adequacy so that losses can be absorbed when such extreme events are realized; these measures are Value-at-Risk and (the somewhat newer) Expected Shortfall. In the first section of our paper we present the motivation behind the development and use of such risk measures, while the concept of "Coherent Risk Measures" is also introduced. In the second section VaR and ES are defined, compared and some of their properties are discussed. An extended section follows describing the various methodologies that often appear in the literature for measuring VaR and ES. The section after that concerns forecast evaluation, which is achieved through the "backtesting" procedure; various tests are presented both for VaR and ES. This concludes the overview of the literature; an empirical application of the concepts discussed comes next: various VaR methodologies (including a novel LASSO-GARCH model) are comparatively studied using a wide array of data sets and (rolling) window sizes (i.e. number of /in sample/ observations). It is demonstrated that the efficacy of all but the most basic forecasting methods depends highly on the data set and the window size used. In addition, our basic implementations of the LASSO-GARCH model seems to outperform the equivalent AR(1)-GARCH models in the vast majority of cases studied and produces results similar to those of Filtered Extreme Value Theory. Closing remarks follow.Τεκμήριο Improved portfolio optimisation with machine learning techniques(03/01/2021) Bletsogiannis, Nikolaos; Μπλετσογιάννης, Νικόλαος; Athens University of Economics and Business, Department of International and European Economic Studies; Tzavalis, Elias; Arvanitis, Stylianos; Dendramis, YiannisThe standard approach for risky asset allocation is the mean-variance framework (Markowitz, 1952). This straightforward approach, of using the sample estimates the inputs, imposes a great amount of error. In this thesis, we implemented different methodologies in order to reduce the estimation error. In Chapter 2 we analyze approaches for a more accurate covariance matrix estimation. These are, the 1-factor model (Sharpe, 1963), shrinkage estimators (Ledoit & Wolf, 2001) and the sparsity principle (Torri, et al., 2018). Also, a brief reference to multivariate GARCH models has been made (Silvennoinen & Teräsvirta, 2008). In Chapter 3 we analyse machine learning models which are used to estimate the conditional expectation of returns. We start with the simple linear regression and continue with shrinkage methods like (Ridge, Lasso and Elastic Net), factor-based approaches like (PCA and PLS). Also, extensive analysis has been made to random forests and neural networks which provide a very accurate expected return estimation (Gu, et al., 2019). Most of the methodologies described above are implemented in real data by estimating the mean-variance (MV) and minimum variance (GMVP) portfolios. The main findings are that shrinkage and sparse precision matrices generate better out of sample returns than the sample minimum variance portfolio in Large N framework. For the mean-variance portfolio, the random forest model generates promising out of sample results but doesn’t outperform the benchmark 1/N approach. In addition, we generated simulated data which then are used to compare the estimation error the sample inverse covariance estimation has with the sparse estimate. The key result is that sparsity can ameliorate the results and provide a much more accurate estimate. Also, in situations where the number of assets is larger than the time series of returns, thus the sample covariance is singular, the sparce approach ends up in a fairly accurate result.