Pub Date : 2023-10-31DOI: 10.3905/jfds.2023.5.4.001
Francesco A. Fabozzi
{"title":"Managing Editor’s Letter","authors":"Francesco A. Fabozzi","doi":"10.3905/jfds.2023.5.4.001","DOIUrl":"https://doi.org/10.3905/jfds.2023.5.4.001","url":null,"abstract":"","PeriodicalId":199045,"journal":{"name":"The Journal of Financial Data Science","volume":"71 5","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135928827","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Randall Davis, Andrew W. Lo, Sudhanshu Mishra, Arash Nourian, Manish Singh, Nicholas Wu, Ruixun Zhang
In this work, the authors create machine learning (ML) models to forecast home equity credit risk for individuals using a real-world dataset and demonstrate methods to explain the output of these ML models to make them more accessible to the end user. They analyze the explainability for various stakeholders: loan companies, regulators, loan applicants, and data scientists, incorporating their different requirements with respect to explanations. For loan companies, they generate explanations for every model prediction of creditworthiness. For regulators, they perform a stress test for extreme scenarios. For loan applicants, they generate diverse counterfactuals to guide them with steps toward a favorable classification from the model. Finally, for data scientists, they generate simple rules that accurately explain 70%–72% of the dataset. Their study provides a synthesized ML explanation framework for all stakeholders and is intended to accelerate the adoption of ML techniques in domains that would benefit from explanations of their predictions.
{"title":"Explainable Machine Learning Models of Consumer Credit Risk","authors":"Randall Davis, Andrew W. Lo, Sudhanshu Mishra, Arash Nourian, Manish Singh, Nicholas Wu, Ruixun Zhang","doi":"10.3905/jfds.2023.1.141","DOIUrl":"https://doi.org/10.3905/jfds.2023.1.141","url":null,"abstract":"In this work, the authors create machine learning (ML) models to forecast home equity credit risk for individuals using a real-world dataset and demonstrate methods to explain the output of these ML models to make them more accessible to the end user. They analyze the explainability for various stakeholders: loan companies, regulators, loan applicants, and data scientists, incorporating their different requirements with respect to explanations. For loan companies, they generate explanations for every model prediction of creditworthiness. For regulators, they perform a stress test for extreme scenarios. For loan applicants, they generate diverse counterfactuals to guide them with steps toward a favorable classification from the model. Finally, for data scientists, they generate simple rules that accurately explain 70%–72% of the dataset. Their study provides a synthesized ML explanation framework for all stakeholders and is intended to accelerate the adoption of ML techniques in domains that would benefit from explanations of their predictions.","PeriodicalId":199045,"journal":{"name":"The Journal of Financial Data Science","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-10-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135592742","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Nusret Cakici, Christian Fieberg, Daniel Metko, Adam Zaremba
Researchers and practitioners hope that machine learning strategies will deliver better performance than traditional methods. But do they? This study documents that stock return predictability with machine learning depends critically on three dimensions: forecast horizon, firm size, and time. It works well for short-term returns, small firms, and early historical data; however, it disappoints in opposite cases. Consequently, annual return forecasts have failed to produce substantial economic gains within most of the US market in the past two decades. These findings challenge the practical utility of predicting returns with machine learning models.
{"title":"Predicting Returns with Machine Learning across Horizons, Firm Size, and Time","authors":"Nusret Cakici, Christian Fieberg, Daniel Metko, Adam Zaremba","doi":"10.3905/jfds.2023.1.139","DOIUrl":"https://doi.org/10.3905/jfds.2023.1.139","url":null,"abstract":"Researchers and practitioners hope that machine learning strategies will deliver better performance than traditional methods. But do they? This study documents that stock return predictability with machine learning depends critically on three dimensions: forecast horizon, firm size, and time. It works well for short-term returns, small firms, and early historical data; however, it disappoints in opposite cases. Consequently, annual return forecasts have failed to produce substantial economic gains within most of the US market in the past two decades. These findings challenge the practical utility of predicting returns with machine learning models.","PeriodicalId":199045,"journal":{"name":"The Journal of Financial Data Science","volume":"235 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135386904","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This article explores artificial neural network (ANN) as a model-free solution for a calibration algorithm of option-pricing models. The authors construct ANNs to calibrate parameters for two well-known GARCH-type option-pricing models: Duan’s GARCH and the classical tempered stable GARCH models that significantly improve upon the limitation of the Black–Scholes model but have suffered from computation complexity. To mitigate this technical difficulty, the authors train ANNs with a dataset generated by the Monte Carlo simulation (MCS) method and apply them to calibrate optimal parameters. The performance results indicate that the ANN approach consistently outperforms MCS and takes advantage of faster computation times once trained. The Greeks of options are also discussed.
{"title":"Deep Calibration with Artificial Neural Network: A Performance Comparison on Option-Pricing Models","authors":"Young Shin Kim, Hyangju Kim, Jaehyung Choi","doi":"10.3905/jfds.2023.1.140","DOIUrl":"https://doi.org/10.3905/jfds.2023.1.140","url":null,"abstract":"This article explores artificial neural network (ANN) as a model-free solution for a calibration algorithm of option-pricing models. The authors construct ANNs to calibrate parameters for two well-known GARCH-type option-pricing models: Duan’s GARCH and the classical tempered stable GARCH models that significantly improve upon the limitation of the Black–Scholes model but have suffered from computation complexity. To mitigate this technical difficulty, the authors train ANNs with a dataset generated by the Monte Carlo simulation (MCS) method and apply them to calibrate optimal parameters. The performance results indicate that the ANN approach consistently outperforms MCS and takes advantage of faster computation times once trained. The Greeks of options are also discussed.","PeriodicalId":199045,"journal":{"name":"The Journal of Financial Data Science","volume":"71 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135426936","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Alik Sokolov, Joshua Kim, Brydon Parker, Benjamin Fattori, Luis Seco
This article introduces a new financial time-series representation model called representations of interrelated financial time series (RIFT). RIFT combines a novel pretraining task and neural network architecture to create generalized representations of multiple financial time-series inputs. The network uses a Siamese architecture to predict pairwise future correlations of securities; the encoder can then be used to create representations of individual securities for downstream tasks. Similar to successful applications of transfer learning in other domains, the authors test the representations on several downstream tasks common in quantitative finance, including dimensionality reduction, portfolio optimization, and portfolio reconstruction. In particular, the article introduces neural hierarchical risk parity (HRP), an improvement on the HRP algorithm, the current state of the art for portfolio optimization, and shows promising results across a variety of assessment criteria, including a 6.0% relative improvement in annualized returns and a 5.6% relative improvement in the Sharpe ratio.
{"title":"RIFT: Pretraining and Applications for Representations of Interrelated Financial Time Series","authors":"Alik Sokolov, Joshua Kim, Brydon Parker, Benjamin Fattori, Luis Seco","doi":"10.3905/jfds.2023.1.138","DOIUrl":"https://doi.org/10.3905/jfds.2023.1.138","url":null,"abstract":"This article introduces a new financial time-series representation model called representations of interrelated financial time series (RIFT). RIFT combines a novel pretraining task and neural network architecture to create generalized representations of multiple financial time-series inputs. The network uses a Siamese architecture to predict pairwise future correlations of securities; the encoder can then be used to create representations of individual securities for downstream tasks. Similar to successful applications of transfer learning in other domains, the authors test the representations on several downstream tasks common in quantitative finance, including dimensionality reduction, portfolio optimization, and portfolio reconstruction. In particular, the article introduces neural hierarchical risk parity (HRP), an improvement on the HRP algorithm, the current state of the art for portfolio optimization, and shows promising results across a variety of assessment criteria, including a 6.0% relative improvement in annualized returns and a 5.6% relative improvement in the Sharpe ratio.","PeriodicalId":199045,"journal":{"name":"The Journal of Financial Data Science","volume":"317 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135966680","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This study leverages novel machine learning techniques to quantify the complex empirical relationship between mortgage default and its drivers. The primary model employed is the authors’ newly developed shape-constrained GAMI-Net, which introduces lattice function-based main effects and pairwise interactions that take user-defined shape constraints. Their approach of adding shape constraints to a lattice module enhances the interpretability and applicability of the model in real-world scenarios. The authors compare the performance of shape-constrained GAMI-Net with alternative machine learning and traditional statistical methods using Freddie Mac’s publicly available mortgage dataset. The results demonstrate competitive predictive performance and high interpretability for the shape-constrained GAMI-Net model.
{"title":"The Anatomy of Mortgage Default Using Shape-Constrained Explainable Machine Learning Model","authors":"Geng Deng, Guangning Xu, Zebin Yang, Yongping Liang, Xindong Wang, Qiang Fu, Aijun Zhang, Agus Sudjianto","doi":"10.3905/jfds.2023.1.136","DOIUrl":"https://doi.org/10.3905/jfds.2023.1.136","url":null,"abstract":"This study leverages novel machine learning techniques to quantify the complex empirical relationship between mortgage default and its drivers. The primary model employed is the authors’ newly developed shape-constrained GAMI-Net, which introduces lattice function-based main effects and pairwise interactions that take user-defined shape constraints. Their approach of adding shape constraints to a lattice module enhances the interpretability and applicability of the model in real-world scenarios. The authors compare the performance of shape-constrained GAMI-Net with alternative machine learning and traditional statistical methods using Freddie Mac’s publicly available mortgage dataset. The results demonstrate competitive predictive performance and high interpretability for the shape-constrained GAMI-Net model.","PeriodicalId":199045,"journal":{"name":"The Journal of Financial Data Science","volume":"137 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135396852","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Risk budgeting (RB) portfolio optimization is one of the popular methods in asset allocation. The key benefit of this method is to control the risk contribution of each asset individually and reduce the unnecessary fluctuation in the allocation by not relying on the expected return of assets. The RB portfolio optimization requires one important parameter, a risk budget vector, and the portfolio performance is strongly influenced by the delicate choice of the values in this vector. Moreover, if the risk strategy allows deviation from a predefined risk budget, then it introduces the problem of finding the optimal time-dependent risk budget deviations. In this article, the author presents a reinforcement learning framework that can select this critical parameter optimally by learning how to control time-dynamic risk budgets in an automated and efficient manner. The experiment result shows that our agent can improve the target performance metric with statistical significance in the different asset universes, indicating that our agent can pick close to optimal risk budget deviations based on the learned policy.
{"title":"Risk Budgeting Portfolio Optimization with Deep Reinforcement Learning","authors":"Seungwoo Han","doi":"10.3905/jfds.2023.1.137","DOIUrl":"https://doi.org/10.3905/jfds.2023.1.137","url":null,"abstract":"Risk budgeting (RB) portfolio optimization is one of the popular methods in asset allocation. The key benefit of this method is to control the risk contribution of each asset individually and reduce the unnecessary fluctuation in the allocation by not relying on the expected return of assets. The RB portfolio optimization requires one important parameter, a risk budget vector, and the portfolio performance is strongly influenced by the delicate choice of the values in this vector. Moreover, if the risk strategy allows deviation from a predefined risk budget, then it introduces the problem of finding the optimal time-dependent risk budget deviations. In this article, the author presents a reinforcement learning framework that can select this critical parameter optimally by learning how to control time-dynamic risk budgets in an automated and efficient manner. The experiment result shows that our agent can improve the target performance metric with statistical significance in the different asset universes, indicating that our agent can pick close to optimal risk budget deviations based on the learned policy.","PeriodicalId":199045,"journal":{"name":"The Journal of Financial Data Science","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135395574","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This article aims to investigate the presence of herding behavior in artificial intelligence (AI)–themed cryptocurrencies following the launch of ChatGPT. The authors analyze daily data from major AI-themed cryptocurrencies between November 2022 and February 2023. This study finds evidence of irrationality among investors in this market segment who tend to imitate others’ decisions regardless of their own beliefs during down events. The authors connect this finding to the herding theory in financial economics and highlight the implications for investors and policymakers. This article contributes to the literature on the impact of AI on financial markets and suggests the need for further research in this area. Finally, this study provides important policy implications, as it could help investors better understand the risks associated with this emerging asset class.
{"title":"Testing for Herding in Artificial Intelligence-Themed Cryptocurrencies Following the Launch of ChatGPT","authors":"Antonis Ballis, Dimitris Anastasiou","doi":"10.3905/jfds.2023.1.134","DOIUrl":"https://doi.org/10.3905/jfds.2023.1.134","url":null,"abstract":"This article aims to investigate the presence of herding behavior in artificial intelligence (AI)–themed cryptocurrencies following the launch of ChatGPT. The authors analyze daily data from major AI-themed cryptocurrencies between November 2022 and February 2023. This study finds evidence of irrationality among investors in this market segment who tend to imitate others’ decisions regardless of their own beliefs during down events. The authors connect this finding to the herding theory in financial economics and highlight the implications for investors and policymakers. This article contributes to the literature on the impact of AI on financial markets and suggests the need for further research in this area. Finally, this study provides important policy implications, as it could help investors better understand the risks associated with this emerging asset class.","PeriodicalId":199045,"journal":{"name":"The Journal of Financial Data Science","volume":"870 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134912805","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
David Blitz, Matthias X. Hanauer, Tobias Hoogteijling, Clint Howard
Machine learning (ML) models for predicting stock returns are typically trained on one-month forward returns. Although these models show impressive full-sample gross alphas, their performance net of transaction costs post-2004 is close to zero. By training on longer prediction horizons and using efficient portfolio construction rules, the authors demonstrate that ML-based investment strategies can still yield significant positive net returns. Longer-horizon strategies select slower signals and load more on traditional asset pricing factors but still unlock unique alpha. The authors conclude that design choices are critical for the success of ML models in real-life applications.
{"title":"The Term Structure of Machine Learning Alpha","authors":"David Blitz, Matthias X. Hanauer, Tobias Hoogteijling, Clint Howard","doi":"10.3905/jfds.2023.1.135","DOIUrl":"https://doi.org/10.3905/jfds.2023.1.135","url":null,"abstract":"Machine learning (ML) models for predicting stock returns are typically trained on one-month forward returns. Although these models show impressive full-sample gross alphas, their performance net of transaction costs post-2004 is close to zero. By training on longer prediction horizons and using efficient portfolio construction rules, the authors demonstrate that ML-based investment strategies can still yield significant positive net returns. Longer-horizon strategies select slower signals and load more on traditional asset pricing factors but still unlock unique alpha. The authors conclude that design choices are critical for the success of ML models in real-life applications.","PeriodicalId":199045,"journal":{"name":"The Journal of Financial Data Science","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-09-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134913113","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：