In the last decade, kernel-based learning approaches typically employed for classification and regression have shown outstanding performance also in dynamic system identification. The typical way to compute the solution of this learning problem subsumes the inversion of the kernel matrix. However, due to limited machine precision, this might not be possible in many practical applications. In this work, we analyze the above problem and show that the typical estimate is just one of the possible infinite solutions that can be leveraged, considering both the supervised and the semi-supervised settings. We show under which conditions the infinite solutions are equivalent and, if it is not the case, we provide a bound on the mismatch between two generic solutions. Then, we propose two specific solutions that are particularly suited to boost sparsity or performance.
(2021). A note on the numerical solutions of kernel-based learning problems [journal article - articolo]. In IEEE TRANSACTIONS ON AUTOMATIC CONTROL. Retrieved from http://hdl.handle.net/10446/171528
A note on the numerical solutions of kernel-based learning problems
Scandella, Matteo;Mazzoleni, Mirko;Formentin, Simone;Previdi, Fabio
2021-01-01
Abstract
In the last decade, kernel-based learning approaches typically employed for classification and regression have shown outstanding performance also in dynamic system identification. The typical way to compute the solution of this learning problem subsumes the inversion of the kernel matrix. However, due to limited machine precision, this might not be possible in many practical applications. In this work, we analyze the above problem and show that the typical estimate is just one of the possible infinite solutions that can be leveraged, considering both the supervised and the semi-supervised settings. We show under which conditions the infinite solutions are equivalent and, if it is not the case, we provide a bound on the mismatch between two generic solutions. Then, we propose two specific solutions that are particularly suited to boost sparsity or performance.File | Dimensione del file | Formato | |
---|---|---|---|
09076872.pdf
Solo gestori di archivio
Versione:
postprint - versione referata/accettata senza referaggio
Licenza:
Licenza default Aisberg
Dimensione del file
768.71 kB
Formato
Adobe PDF
|
768.71 kB | Adobe PDF | Visualizza/Apri |
Pubblicazioni consigliate
Aisberg ©2008 Servizi bibliotecari, Università degli studi di Bergamo | Terms of use/Condizioni di utilizzo