Hyperparameter Optimization for Portfolio Selection
Nystrup, Peter LU
; Lindström, Erik
LU
and Henrik, Madsen
(2020)
In The Journal of Financial Data Science
2(2).
- Abstract
- Portfolio selection involves a trade-off between maximizing expected return and minimizing risk. In practice, useful formulations also include various costs and constraints that regularize the problem and reduce the risk due to estimation errors, resulting in solutions that depend on a number of hyperparameters. As the number of hyperparameters grows, selecting their value becomes increasingly important and difficult. In this article, the authors propose a systematic approach to hyperparameter optimization by leveraging recent advances in automated machine learning and multiobjective optimization. They optimize hyperparameters on a train set to yield the best result subject to market-determined realized costs. In applications to single-... (More)
- Portfolio selection involves a trade-off between maximizing expected return and minimizing risk. In practice, useful formulations also include various costs and constraints that regularize the problem and reduce the risk due to estimation errors, resulting in solutions that depend on a number of hyperparameters. As the number of hyperparameters grows, selecting their value becomes increasingly important and difficult. In this article, the authors propose a systematic approach to hyperparameter optimization by leveraging recent advances in automated machine learning and multiobjective optimization. They optimize hyperparameters on a train set to yield the best result subject to market-determined realized costs. In applications to single- and multiperiod portfolio selection, they show that sequential hyperparameter optimization finds solutions with better risk–return trade-offs than manual, grid, and random search over hyperparameters using fewer function evaluations. At the same time, the solutions found are more stable from in-sample training to out-of-sample testing, suggesting they are less likely to be extremities that randomly happened to yield good performance in training. (Less)
- Abstract (Swedish)
- Portfolio optimization involves a tradeoff between maximizing expected return
and minimizing risk. In practice, useful formulations also include various costs and constraints that regularize the problem and reduce the risk due to estimation errors, resulting in solutions that depend on a number of hyperparameters. As the number of hyperparameters grows, selecting their value becomes increasingly important and difficult. In this article we propose a systematic approach to hyperparameter optimization by leveraging recent advances in automated machine learning and multi-objective optimization. We optimize hyperparameters on a train set to yield the best result subject to market-determined realized costs. For example, we show that when the... (More) - Portfolio optimization involves a tradeoff between maximizing expected return
and minimizing risk. In practice, useful formulations also include various costs and constraints that regularize the problem and reduce the risk due to estimation errors, resulting in solutions that depend on a number of hyperparameters. As the number of hyperparameters grows, selecting their value becomes increasingly important and difficult. In this article we propose a systematic approach to hyperparameter optimization by leveraging recent advances in automated machine learning and multi-objective optimization. We optimize hyperparameters on a train set to yield the best result subject to market-determined realized costs. For example, we show that when the signal-to-noise ratio of return forecasts deteriorates, the optimal level of transaction costs imposed in portfolio optimization increases to prevent excessive
noise trading. In applications to single- and multi-period portfolio optimization,
we show that sequential hyperparameter optimization finds solutions with better
return/risk tradeoffs than manual, grid, and random search over hyperparameters using fewer function evaluations. At the same time, the solutions found are more stable from in-sample training to out-of-sample testing, suggesting they are less likely to be extremities that randomly happened to yield good performance in training. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/b8e37d41-3a29-4ef4-9a2e-89f57a3ff78d
- author
- Nystrup, Peter
LU
; Lindström, Erik
LU
and Henrik, Madsen
- organization
- publishing date
- 2020-06-18
- type
- Contribution to journal
- publication status
- published
- subject
- in
- The Journal of Financial Data Science
- volume
- 2
- issue
- 2
- publisher
- Portfolio Management Research
- external identifiers
-
- scopus:85110417315
- ISSN
- 2640-3943
- DOI
- 10.3905/jfds.2020.1.035
- language
- English
- LU publication?
- yes
- id
- b8e37d41-3a29-4ef4-9a2e-89f57a3ff78d
- date added to LUP
- 2020-03-17 10:04:30
- date last changed
- 2023-04-27 12:35:28
@article{b8e37d41-3a29-4ef4-9a2e-89f57a3ff78d,
abstract = {{Portfolio selection involves a trade-off between maximizing expected return and minimizing risk. In practice, useful formulations also include various costs and constraints that regularize the problem and reduce the risk due to estimation errors, resulting in solutions that depend on a number of hyperparameters. As the number of hyperparameters grows, selecting their value becomes increasingly important and difficult. In this article, the authors propose a systematic approach to hyperparameter optimization by leveraging recent advances in automated machine learning and multiobjective optimization. They optimize hyperparameters on a train set to yield the best result subject to market-determined realized costs. In applications to single- and multiperiod portfolio selection, they show that sequential hyperparameter optimization finds solutions with better risk–return trade-offs than manual, grid, and random search over hyperparameters using fewer function evaluations. At the same time, the solutions found are more stable from in-sample training to out-of-sample testing, suggesting they are less likely to be extremities that randomly happened to yield good performance in training.}},
author = {{Nystrup, Peter and Lindström, Erik and Henrik, Madsen}},
issn = {{2640-3943}},
language = {{eng}},
month = {{06}},
number = {{2}},
publisher = {{Portfolio Management Research}},
series = {{The Journal of Financial Data Science}},
title = {{Hyperparameter Optimization for Portfolio Selection}},
url = {{http://dx.doi.org/10.3905/jfds.2020.1.035}},
doi = {{10.3905/jfds.2020.1.035}},
volume = {{2}},
year = {{2020}},
}