Explainable artificial intelligence for crypto asset allocation

Babaei, Golnoosh; Giudici, Paolo; Raffinetti, Emanuela

doi:10.1016/j.frl.2022.102941

Cited by 31 publications

(8 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, our study provides a pioneering step towards a deeper understanding of XAI as discussed in De Bock 2023 and we suggest that the presented framework can be applied to relevant quantitative approaches in financial management with metric target variables. From our point of view, we consider an assessment of XAI applied to credit risk modelling, as discussed in Kellner et al (2022), multivariate time series forecasts, as presented in Ahelegbey et al (2016) and Giudici et al (2020), as well as portfolio management, as introduced in Babaei et al (2022), a promising next step towards a deeper understanding of XAI in finance.…”

Section: Discussionmentioning

confidence: 99%

“…Despite the promising performance of ML techniques in the context of economic data describes a growing field of research, the lack of explainability in the context of financial data is not adequately discussed, yet. Taking into account that financial applications such as return predictions (see Avramov et al 2022), asset management (see Babaei et al 2022) or credit risk assessment (see Kellner et al 2022) are highly relevant for financial institutions, the lack of interpretability of ML techniques presents a crucial challenge for regulatory demands. The challenge being that regulatory demands require interpretability as a necessary precondition.…”

Section: Introductionmentioning

confidence: 99%

“…The challenge being that regulatory demands require interpretability as a necessary precondition. In this vein, Giudici and Raffinetti (2022) address cyber risk management, Babaei et al (2023) the lending process within a fintech universe for small end medium enterprises and Babaei et al (2022) provide a discussion on asset allocation and crypto assets. 3 Up to now, recent literature on ML is mainly focused on classification problems and the assessment of individual feature contributions, especially when the target variable of the assessment is measured in a metric scale, is of growing interest not only in a financial context (see De Bock 2023.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

On the information content of explainable artificial intelligence for quantitative approaches in finance

Berger

2024

OR Spectrum

View full text Add to dashboard Cite

We simulate economic data to apply state-of-the-art machine learning algorithms and analyze the economic precision of competing concepts for model agnostic explainable artificial intelligence (XAI) techniques. Also, we assess empirical data and provide a discussion of the competing approaches in comparison with econometric benchmarks, when the data-generating process is unknown. The simulation assessment provides evidence that the applied XAI techniques provide similar economic information on relevant determinants when the data generating process is linear. We find that the adequate choice of XAI technique is crucial when the data generating process is unknown. In comparison to econometric benchmark models, the application of boosted regression trees in combination with Shapley values combines both a superior fit to the data and innovative interpretable insights into non-linear impact factors. Therefore it describes a promising alternative to the econometric benchmark approach.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

On the information content of explainable artificial intelligence for quantitative approaches in finance

Berger

2024

OR Spectrum

View full text Add to dashboard Cite

show abstract

“…Barbaglia et al [25] use ALE plots to determine the drivers of mortgage probability of defaults in Europe. In related fields, such as cyber risk management or financial risk management in general, the application of XAI methods becomes more widespread as well, see, e.g., [62][63][64][65].…”

Section: Methodsmentioning

confidence: 99%

Quantifying uncertainty of machine learning methods for loss given default

Nagl

Rösch

2022

Front. Appl. Math. Stat.

View full text Add to dashboard Cite

Machine learning has increasingly found its way into the credit risk literature. When applied to forecasting credit risk parameters, the approaches have been found to outperform standard statistical models. The quantification of prediction uncertainty is typically not analyzed in the machine learning credit risk setting. However, this is vital to the interests of risk managers and regulators alike as its quantification increases the transparency and stability in risk management and reporting tasks. We fill this gap by applying the novel approach of deep evidential regression to loss given defaults (LGDs). We evaluate aleatoric and epistemic uncertainty for LGD estimation techniques and apply explainable artificial intelligence (XAI) methods to analyze the main drivers. We find that aleatoric uncertainty is considerably larger than epistemic uncertainty. Hence, the majority of uncertainty in LGD estimates appears to be irreducible as it stems from the data itself.

show abstract

“…As the name implies, blockchain is a decentralized technological platform for storing and managing data and transactions without the need for third parties (Yli-Huumo et al 2016), and cryptocurrency refers to tokens or digital currency generated from cryptography Ghaemi Asl and Roubaud Financial Innovation (2024) 10:89 Various studies have attempted to find evidence of a nexus between the cryptocurrency market and AI using different technical methods. By combining random forest models with Shapley values for predicting cryptographic assets, Babaei et al (2022) implemented a methodology that achieved predictability and explainability in the allocation of AI-based cryptographic assets. According to Cho et al (2021), there has been a growing body of research related to financial markets that apply machine learning to cryptocurrencies using blockchain technology.…”

Section: Literature Reviewmentioning

confidence: 99%

Asymmetric interactions among cutting-edge technologies and pioneering conventional and Islamic cryptocurrencies: fresh evidence from intra-day-based good and bad volatilities

Ghaemi Asl,

Roubaud

2024

Financ Innov

View full text Add to dashboard Cite

This study examines the nexus between the good and bad volatilities of three technological revolutions—financial technology (FinTech), the Internet of Things, and artificial intelligence and technology—as well as the two main conventional and Islamic cryptocurrency platforms, Bitcoin and Stellar, via three approaches: quantile cross-spectral coherence, quantile-VAR connectedness, and quantile-based non-linear causality-in-mean and variance analysis. The results are as follows: (1) under normal market conditions, in long-run horizons there is a significant positive cross-spectral relationship between FinTech's positive volatilities and Stellar’s negative volatilities; (2) Stellar’s negative and positive volatilities exhibit the highest net spillovers at the lower and upper tails, respectively; and (3) the quantile-based causality results indicate that Bitcoin’s good (bad) volatilities can lead to bad (good) volatilities in all three smart technologies operating between normal and bull market conditions. Moreover, the Bitcoin industry’s negative volatilities have a bilateral cause-and-effect relationship with FinTech’s positive volatilities. By analyzing the second moment, we found that Bitcoin's negative volatilities are the only cause variable that generates FinTech's good volatility in a unidirectional manner. As for Stellar, only bad volatilities have the potential to signal good volatilities for cutting-edge technologies in some middle quantiles, whereas good volatilities have no significant effect. Hence, the trade-off between Bitcoin and cutting-edge technologies, especially FinTech-related advancements, appear more broadly and randomly compared with the Stellar-innovative technologies nexus. The findings provide valuable insights for FinTech companies, blockchain developers, crypto-asset regulators, portfolio managers, and high-tech investors.

show abstract

Explainable artificial intelligence for crypto asset allocation

Cited by 31 publications

References 17 publications

On the information content of explainable artificial intelligence for quantitative approaches in finance

On the information content of explainable artificial intelligence for quantitative approaches in finance

Quantifying uncertainty of machine learning methods for loss given default

Asymmetric interactions among cutting-edge technologies and pioneering conventional and Islamic cryptocurrencies: fresh evidence from intra-day-based good and bad volatilities

Contact Info

Product

Resources

About