Generalized Exponentiated Gradient Algorithms and Their Application to On-Line Portfolio Selection

Stochastic gradient descent (SGD) and exponentiated gradient (EG) update methods are widely used in signal processing and machine learning. This study introduces a novel family of generalized Exponentiated Gradient updates (EGAB) derived from the alpha-beta (AB) divergence regularization. The EGAB f...

Full description

Saved in:
Bibliographic Details
Main Authors: Andrzej Cichocki, Sergio Cruces, Auxiliadora Sarmiento, Toshihisa Tanaka
Format: Article
Language:English
Published: IEEE 2024-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10807168/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846099863404544000
author Andrzej Cichocki
Sergio Cruces
Auxiliadora Sarmiento
Toshihisa Tanaka
author_facet Andrzej Cichocki
Sergio Cruces
Auxiliadora Sarmiento
Toshihisa Tanaka
author_sort Andrzej Cichocki
collection DOAJ
description Stochastic gradient descent (SGD) and exponentiated gradient (EG) update methods are widely used in signal processing and machine learning. This study introduces a novel family of generalized Exponentiated Gradient updates (EGAB) derived from the alpha-beta (AB) divergence regularization. The EGAB framework provides enhanced flexibility for processing data with varying distributions, thanks to the tunable hyperparameters of the AB divergence. We explore the applicability of these updates in online portfolio selection (OLPS) for financial markets with the goal of developing algorithms that achieve high risk-adjusted returns, even under relatively high transaction costs. The proposed EGAB algorithms are developed using constrained gradient optimization with regularization terms, demonstrating their versatility in OLPS by unifying the directional search of various algorithms and enabling interpolation between them. Our analysis and extensive computer simulations reveal that EGAB updates outperform existing OLPS algorithms, delivering good results on several performance metrics, such as cumulative return, average excess return, Sharpe ratio, and Calmar ratio, especially when transaction costs are significant. In conclusion, this study introduces a new family of exponentiated gradient updates and demonstrates their flexibility and effectiveness through extensive simulations across a wide range of real-world financial datasets.
format Article
id doaj-art-07c770f19e6f4d1d8de43b9071d30e78
institution Kabale University
issn 2169-3536
language English
publishDate 2024-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-07c770f19e6f4d1d8de43b9071d30e782024-12-31T00:01:02ZengIEEEIEEE Access2169-35362024-01-011219700019702010.1109/ACCESS.2024.352038910807168Generalized Exponentiated Gradient Algorithms and Their Application to On-Line Portfolio SelectionAndrzej Cichocki0Sergio Cruces1https://orcid.org/0000-0003-4121-7137Auxiliadora Sarmiento2https://orcid.org/0000-0003-2587-1382Toshihisa Tanaka3https://orcid.org/0000-0002-5056-9508Polish Academy of Science, Systems Research Institute, Warszawa, PolandDepartamento de Teoría de la Señal y Comunicaciones, Universidad de Sevilla, Seville, SpainDepartamento de Teoría de la Señal y Comunicaciones, Universidad de Sevilla, Seville, SpainDepartment of Electrical Engineering and Computer Science, Tokyo University of Agriculture and Technology, Koganei-shi, Tokyo, JapanStochastic gradient descent (SGD) and exponentiated gradient (EG) update methods are widely used in signal processing and machine learning. This study introduces a novel family of generalized Exponentiated Gradient updates (EGAB) derived from the alpha-beta (AB) divergence regularization. The EGAB framework provides enhanced flexibility for processing data with varying distributions, thanks to the tunable hyperparameters of the AB divergence. We explore the applicability of these updates in online portfolio selection (OLPS) for financial markets with the goal of developing algorithms that achieve high risk-adjusted returns, even under relatively high transaction costs. The proposed EGAB algorithms are developed using constrained gradient optimization with regularization terms, demonstrating their versatility in OLPS by unifying the directional search of various algorithms and enabling interpolation between them. Our analysis and extensive computer simulations reveal that EGAB updates outperform existing OLPS algorithms, delivering good results on several performance metrics, such as cumulative return, average excess return, Sharpe ratio, and Calmar ratio, especially when transaction costs are significant. In conclusion, this study introduces a new family of exponentiated gradient updates and demonstrates their flexibility and effectiveness through extensive simulations across a wide range of real-world financial datasets.https://ieeexplore.ieee.org/document/10807168/Alpha-beta divergencesexponentiated gradient algorithmson-line portfolio selection
spellingShingle Andrzej Cichocki
Sergio Cruces
Auxiliadora Sarmiento
Toshihisa Tanaka
Generalized Exponentiated Gradient Algorithms and Their Application to On-Line Portfolio Selection
IEEE Access
Alpha-beta divergences
exponentiated gradient algorithms
on-line portfolio selection
title Generalized Exponentiated Gradient Algorithms and Their Application to On-Line Portfolio Selection
title_full Generalized Exponentiated Gradient Algorithms and Their Application to On-Line Portfolio Selection
title_fullStr Generalized Exponentiated Gradient Algorithms and Their Application to On-Line Portfolio Selection
title_full_unstemmed Generalized Exponentiated Gradient Algorithms and Their Application to On-Line Portfolio Selection
title_short Generalized Exponentiated Gradient Algorithms and Their Application to On-Line Portfolio Selection
title_sort generalized exponentiated gradient algorithms and their application to on line portfolio selection
topic Alpha-beta divergences
exponentiated gradient algorithms
on-line portfolio selection
url https://ieeexplore.ieee.org/document/10807168/
work_keys_str_mv AT andrzejcichocki generalizedexponentiatedgradientalgorithmsandtheirapplicationtoonlineportfolioselection
AT sergiocruces generalizedexponentiatedgradientalgorithmsandtheirapplicationtoonlineportfolioselection
AT auxiliadorasarmiento generalizedexponentiatedgradientalgorithmsandtheirapplicationtoonlineportfolioselection
AT toshihisatanaka generalizedexponentiatedgradientalgorithmsandtheirapplicationtoonlineportfolioselection