RECORD
Portfolio Optimization using Predictive Auxiliary Classifier Generative Adversarial Networks with Measuring Uncertainty
Portfolio Management · Submitted 24 Apr 2023
abstract compiled · 2026-09-23 · no language-model review
The authors' main claim As a result, a portfolio using PredACGAN exhibits 9.123% yearly returns and a Sharpe ratio of 1.054, while a portfolio without considering risk measures shows 1.024% yearly returns and a Sharpe ratio of 0.236 in the same scenario.
The compiled record
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
Open the compiled record
The abstract, sentence by sentence
In financial engineering, portfolio optimization has been of consistent interest. Portfolio optimization is a process of modulating asset distributions to maximize expected returns and minimize risks. To obtain the expected returns, deep learning models have been explored in recent years. However, due to the deterministic nature of the models, it is difficult to consider the risk of portfolios because conventional deep learning models do not know how reliable their predictions can be. To address this limitation, this paper proposes a probabilistic model, namely predictive auxiliary classifier generative adversarial networks (PredACGAN). The proposed PredACGAN utilizes the characteristic of the ACGAN framework in which the output of the generator forms a distribution. While ACGAN has not been employed for predictive models and is generally utilized for image sample generation, this paper proposes a method to use the ACGAN structure for a probabilistic and predictive model. Additionally, an algorithm to use the risk measurement obtained by PredACGAN is proposed. In the algorithm, the assets that are predicted to be at high risk are eliminated from the investment universe at the rebalancing moment. Therefore, PredACGAN considers both return and risk to optimize portfolios. The proposed algorithm and PredACGAN have been evaluated with daily close price data of S&P 500 from 1990 to 2020. Experimental scenarios are assumed to rebalance the portfolios monthly according to predictions and risk measures with PredACGAN. As a result, a portfolio using PredACGAN exhibits 9.123% yearly returns and a Sharpe ratio of 1.054, while a portfolio without considering risk measures shows 1.024% yearly returns and a Sharpe ratio of 0.236 in the same scenario. Also, the maximum drawdown of the proposed portfolio is lower than the portfolio without PredACGAN.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | In financial engineering, portfolio optimization has been of consistent interest. | background 1.00 | 0.03 | 0.01 | asserted 1.00 |
| s1 | Portfolio optimization is a process of modulating asset distributions to maximize expected returns and minimize risks. | background 1.00 | 0.04 | 0.02 | asserted 1.00 |
| s2 | To obtain the expected returns, deep learning models have been explored in recent years. | background 1.00 | 0.05 | 0.01 | asserted 0.99 |
| s3 | However, due to the deterministic nature of the models, it is difficult to consider the risk of portfolios because conventional deep learning models do not know how reliable their predictions can be. | gap 0.94 | 0.07 | 0.02 | asserted 0.99 |
| s4 | To address this limitation, this paper proposes a probabilistic model, namely predictive auxiliary classifier generative adversarial networks (PredACGAN). | contribution 0.97 | 0.88 | 0.08 | asserted 1.00 |
| s5 | The proposed PredACGAN utilizes the characteristic of the ACGAN framework in which the output of the generator forms a distribution. | method 0.96 | 0.29 | 0.03 | asserted 1.00 |
| s6 | While ACGAN has not been employed for predictive models and is generally utilized for image sample generation, this paper proposes a method to use the ACGAN structure for a probabilistic and predictive model. | contribution 0.84 | 0.70 | 0.04 | asserted 1.00 |
| s7 | Additionally, an algorithm to use the risk measurement obtained by PredACGAN is proposed. | contribution 0.70 | 0.90 | 0.04 | asserted 1.00 |
| s8 | In the algorithm, the assets that are predicted to be at high risk are eliminated from the investment universe at the rebalancing moment. | method 0.98 | 0.75 | 0.04 | asserted 1.00 |
| s9 | Therefore, PredACGAN considers both return and risk to optimize portfolios. | contribution 0.50 | 0.56 | 0.25 | asserted 1.00 |
| s10 | The proposed algorithm and PredACGAN have been evaluated with daily close price data of S&P 500 from 1990 to 2020. | method 0.99 | 0.94 | 0.03 | asserted 1.00 |
| s11 | Experimental scenarios are assumed to rebalance the portfolios monthly according to predictions and risk measures with PredACGAN. | method 0.99 | 0.74 | 0.02 | asserted 0.91 |
| s12 | As a result, a portfolio using PredACGAN exhibits 9.123% yearly returns and a Sharpe ratio of 1.054, while a portfolio without considering risk measures shows 1.024% yearly returns and a Sharpe ratio of 0.236 in the same scenario. | result 1.00 | 0.95 | 0.62 | asserted 1.00 |
| s13 | Also, the maximum drawdown of the proposed portfolio is lower than the portfolio without PredACGAN. | result 1.00 | 0.94 | 0.23 | asserted 1.00 |
What was examined, and what was not
- arXiv listing abstract v1
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
The paper
Original abstract
In financial engineering, portfolio optimization has been of consistent interest. Portfolio optimization is a process of modulating asset distributions to maximize expected returns and minimize risks. To obtain the expected returns, deep learning models have been explored in recent years. However, due to the deterministic nature of the models, it is difficult to consider the risk of portfolios because conventional deep learning models do not know how reliable their predictions can be. To address this limitation, this paper proposes a probabilistic model, namely predictive auxiliary classifier generative adversarial networks (PredACGAN). The proposed PredACGAN utilizes the characteristic of the ACGAN framework in which the output of the generator forms a distribution. While ACGAN has not been employed for predictive models and is generally utilized for image sample generation, this paper proposes a method to use the ACGAN structure for a probabilistic and predictive model. Additionally, an algorithm to use the risk measurement obtained by PredACGAN is proposed. In the algorithm, the assets that are predicted to be at high risk are eliminated from the investment universe at the rebalancing moment. Therefore, PredACGAN considers both return and risk to optimize portfolios. The proposed algorithm and PredACGAN have been evaluated with daily close price data of S&P 500 from 1990 to 2020. Experimental scenarios are assumed to rebalance the portfolios monthly according to predictions and risk measures with PredACGAN. As a result, a portfolio using PredACGAN exhibits 9.123% yearly returns and a Sharpe ratio of 1.054, while a portfolio without considering risk measures shows 1.024% yearly returns and a Sharpe ratio of 0.236 in the same scenario. Also, the maximum drawdown of the proposed portfolio is lower than the portfolio without PredACGAN.
Human review
How the literature uses it
What this paper does with the arXiv papers it cites
What it cites · 4 sentences without a confident label
Uses the method or tool of 2 papers
-
The hinge loss [ref] is used for $D$, and Wasserstein distance [ref] is used for $G$, which correspond to $L_r$ and $L_f$ in Eqs. [ref] and [ref].
Geometric GAN · Training details of PredACGAN and experimental scenarios · 1.00
-
In TTUR, Adam optimizer [ref] is used with hyperparameters of $β_1=0.0$ and $β_2=0.9$, and learning rate for $D$ and $G$ are set to $2×10^-5$ and $1×10^-5$, respectively.
Adam: A Method for Stochastic Optimization · Training details of PredACGAN and experimental scenarios · 1.00
Cites as background 1 paper
-
In this manner, GAN has been generally used for synthetic sample generation since the model can produce realistic samples [ref] .
Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks · Auxiliary classifier generative adversarial networks · 0.85
Discussion
Know something about this paper? Leave a comment, or highlight any passage above to comment on it.