Between Nonlinearities, Complexity, and Noises: An Application on Portfolio Selection Using Kernel Principal Component Analysis
Autor: | Igor Ferreira do Nascimento, Yaohao Peng, João Victor Freitas Machado, Pedro Henrique Melo Albuquerque |
---|---|
Jazyk: | angličtina |
Rok vydání: | 2019 |
Předmět: |
covariance estimation
General Physics and Astronomy lcsh:Astrophysics Statistics::Other Statistics 02 engineering and technology random matrix theory Kernel principal component analysis Article Estimation of covariance matrices kernel methods 0502 economics and business lcsh:QB460-466 0202 electrical engineering electronic engineering information engineering Applied mathematics lcsh:Science Eigenvalues and eigenvectors Mathematics 050208 finance Covariance matrix 05 social sciences Estimator nonlinearity portfolio allocation Covariance lcsh:QC1-999 regularization Kernel method machine learning 020201 artificial intelligence & image processing lcsh:Q Random matrix lcsh:Physics high dimensionality |
Zdroj: | Entropy, Vol 21, Iss 4, p 376 (2019) Entropy Volume 21 Issue 4 |
ISSN: | 1099-4300 |
Popis: | This paper discusses the effects of introducing nonlinear interactions and noise-filtering to the covariance matrix used in Markowitz&rsquo s portfolio allocation model, evaluating the technique&rsquo s performances for daily data from seven financial markets between January 2000 and August 2018. We estimated the covariance matrix by applying Kernel functions, and applied filtering following the theoretical distribution of the eigenvalues based on the Random Matrix Theory. The results were compared with the traditional linear Pearson estimator and robust estimation methods for covariance matrices. The results showed that noise-filtering yielded portfolios with significantly larger risk-adjusted profitability than its non-filtered counterpart for almost half of the tested cases. Moreover, we analyzed the improvements and setbacks of the nonlinear approaches over linear ones, discussing in which circumstances the additional complexity of nonlinear features seemed to predominantly add more noise or predictive performance. |
Databáze: | OpenAIRE |
Externí odkaz: | |
Nepřihlášeným uživatelům se plný text nezobrazuje | K zobrazení výsledku je třeba se přihlásit. |