Fumin Zhu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Fumin Zhu has a total of 43 co-authors. Here are the top co-authors by number of shared publications:
- Zunxin Zheng
College of Economics, Shenzhen University, Shenzhen, China
5 shared publications - Chen Zhang
College of Economics, Shenzhen University, Shenzhen, China
2 shared publications - Jinming Wen
Jilin University
2 shared publications - Young Shin Kim
Is an associate professor of finance in the College of Business at Stony Brook University in Stony Brook, NY
2 shared publications Zunxin Zheng
College of Economics Shenzhen University Shenzhen Guangdong Province China
2 shared publications- Ahmed Farouk
College of Science and Engineering, Qatar Center for Quantum Computing, Hamad Bin Khalifa University, Doha, Qatar
1 shared publication
Latest Publications
- 12 September 2026
- Published by Elsevier in The Quarterly Review of Economics and Finance
- Vol. 110
Geopolitical risk and economic policy uncertainty aggravate the vulnerability of crude oil volatility
- 12 September 2026
- Published by Elsevier in The Quarterly Review of Economics and Finance
- Vol. 110
Geopolitical risk and economic policy uncertainty aggravate the vulnerability of crude oil volatility
- Top Cited Papers
- 13 January 2025
- Published by Wiley in International Journal of Finance & Economics
- Vol. 30 (4) , 3856-3874
The Disclosure Fog: Institutional Investors and Corporate Greenwashing
Corporate environmental, social, and governance (ESG) disclosure is often a superficial signal rather than something of substance and is an easily negligible form of greenwashing. Here, we explore the relationship between corporate disclosure greenwashing and institutional investors using data from Chinese listed heavily polluting firms from 2012 to 2021. We hypothesise and discover that while institutional investors encourage firms to publish social responsibility reports, they may actually discourage the use of sustainability‐related phrases in these reports. The findings hold even after performing several endogeneity and robustness tests. We also identify three mechanisms through which institutional investors influence corporate greenwashing: corporate governance, information transparency, and corporate operations. Additionally, pressure‐sensitive institutional investors are more likely to facilitate corporate greenwashing than pressure‐insensitive and pressure‐uncertain investors. However, after the Guidance on Building Green Financial System's implementation in 2016, the positive relationship between disclosure greenwashing and institutional investors has rapidly declined. The promotional effect of institutional investor‐induced greenwashing is also lower in state‐owned enterprises. Overall, our findings not only enrich the literature on institutional investors and corporate information disclosure but also have implications for improving the green financial system.
- 13 January 2025
- Published by Wiley in International Journal of Finance & Economics
- Vol. 30 (4) , 3856-3874
Top Cited PapersThe Disclosure Fog: Institutional Investors and Corporate Greenwashing
Corporate environmental, social, and governance (ESG) disclosure is often a superficial signal rather than something of substance and is an easily negligible form of greenwashing. Here, we explore the relationship between corporate disclosure greenwashing and institutional investors using data from Chinese listed heavily polluting firms from 2012 to 2021. We hypothesise and discover that while institutional investors encourage firms to publish social responsibility reports, they may actually discourage the use of sustainability‐related phrases in these reports. The findings hold even after performing several endogeneity and robustness tests. We also identify three mechanisms through which institutional investors influence corporate greenwashing: corporate governance, information transparency, and corporate operations. Additionally, pressure‐sensitive institutional investors are more likely to facilitate corporate greenwashing than pressure‐insensitive and pressure‐uncertain investors. However, after the Guidance on Building Green Financial System's implementation in 2016, the positive relationship between disclosure greenwashing and institutional investors has rapidly declined. The promotional effect of institutional investor‐induced greenwashing is also lower in state‐owned enterprises. Overall, our findings not only enrich the literature on institutional investors and corporate information disclosure but also have implications for improving the green financial system.
- 24 May 2024
- Published by Elsevier in The North American Journal of Economics and Finance
- Vol. 74
Pricing VIX options based on mean-reverting models driven by information
- 24 May 2024
- Published by Elsevier in The North American Journal of Economics and Finance
- Vol. 74
Pricing VIX options based on mean-reverting models driven by information
- 12 March 2024
- Published by Wiley in European Financial Management
- Vol. 30 (5) , 2596-2627
Bond return predictability: Macro factors and machine learning methods
We investigate the impact of macroeconomic variables on bond risk premia prediction via machine learning techniques. On the basis of Chinese treasury bonds from March 2006 to December 2022, we show that adding macroeconomic factors improves bond return forecasts and generates higher economic benefits to investors. This is achieved when the nonlinear relationship between macroeconomic variables and bond returns is modelled via machine learning methods. Furthermore, the importance of macroeconomic determinants changes along the yield curve. Our study sheds new light on the information contained in macroeconomic variables for treasury bond valuation and highlights the importance of utilizing appropriate machine learning methods.
- 12 March 2024
- Published by Wiley in European Financial Management
- Vol. 30 (5) , 2596-2627
Bond return predictability: Macro factors and machine learning methods
We investigate the impact of macroeconomic variables on bond risk premia prediction via machine learning techniques. On the basis of Chinese treasury bonds from March 2006 to December 2022, we show that adding macroeconomic factors improves bond return forecasts and generates higher economic benefits to investors. This is achieved when the nonlinear relationship between macroeconomic variables and bond returns is modelled via machine learning methods. Furthermore, the importance of macroeconomic determinants changes along the yield curve. Our study sheds new light on the information contained in macroeconomic variables for treasury bond valuation and highlights the importance of utilizing appropriate machine learning methods.
- 18 May 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Transportation Systems
- Vol. 24 (7) , 7621-7630
A Pseudo-Inverse-Based Hard Thresholding Algorithm for Sparse Signal Recovery
Acquiring a sparse signal from an underdetermined linear system arises from numerous applications. Several hard thresholding algorithms have been developed for the sparse reconstruction. In particular, the so-called Newton-Step-based Iterative Hard Thresholding (NSIHT) and Newton-Step-based Hard Thresholding Pursuit (NSHTP) algorithms were developed by Meng et al. recently. Although they are efficient, a faster sparse recovery algorithm with a better reconstruction performance is still needed. In this paper, we first propose a Pseudo-inverse-based Hard Thresholding sparse signal recovery algorithm called PHT for short. Unlike the Iterative Hard Thresholding (IHT) algorithm which uses the gradient of the objective function to iteratively update the solution, our proposed algorithm utilizes the pseudo-inverse of the sensing matrix to iteratively update the solution. We then analyze the computational complexity of PHT and show that it is O(n²/m²) times faster than both NSIHT and NSHTP if they perform the same number of iterations, where m and n are the number of rows and columns of the sensing matrix \A. Furthermore, we establish a sufficient condition of stable recovery of the sparse signal with PHT by using the restricted isometry property (RIP) of the sensing matrix. Finally, extensive experiments are conducted which indicate that our proposed algorithm PHT is much faster than both NSIHT and NSHTP with overall better recovery performance.
- 18 May 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Transportation Systems
- Vol. 24 (7) , 7621-7630
A Pseudo-Inverse-Based Hard Thresholding Algorithm for Sparse Signal Recovery
Acquiring a sparse signal from an underdetermined linear system arises from numerous applications. Several hard thresholding algorithms have been developed for the sparse reconstruction. In particular, the so-called Newton-Step-based Iterative Hard Thresholding (NSIHT) and Newton-Step-based Hard Thresholding Pursuit (NSHTP) algorithms were developed by Meng et al. recently. Although they are efficient, a faster sparse recovery algorithm with a better reconstruction performance is still needed. In this paper, we first propose a Pseudo-inverse-based Hard Thresholding sparse signal recovery algorithm called PHT for short. Unlike the Iterative Hard Thresholding (IHT) algorithm which uses the gradient of the objective function to iteratively update the solution, our proposed algorithm utilizes the pseudo-inverse of the sensing matrix to iteratively update the solution. We then analyze the computational complexity of PHT and show that it is O(n²/m²) times faster than both NSIHT and NSHTP if they perform the same number of iterations, where m and n are the number of rows and columns of the sensing matrix \A. Furthermore, we establish a sufficient condition of stable recovery of the sparse signal with PHT by using the restricted isometry property (RIP) of the sensing matrix. Finally, extensive experiments are conducted which indicate that our proposed algorithm PHT is much faster than both NSIHT and NSHTP with overall better recovery performance.
- 1 October 2021
- Published by Elsevier in Computers and Electrical Engineering
- Vol. 95, 107373
Potential technologies and applications based on deep learning in the 6G networks
- 1 October 2021
- Published by Elsevier in Computers and Electrical Engineering
- Vol. 95, 107373
Potential technologies and applications based on deep learning in the 6G networks
- 29 September 2021
- Published by Springer Nature in Mathematical Methods of Operations Research
- Vol. 94 (2) , 197-217
Aumann–Serrano index of risk in portfolio optimization
- 29 September 2021
- Published by Springer Nature in Mathematical Methods of Operations Research
- Vol. 94 (2) , 197-217
Aumann–Serrano index of risk in portfolio optimization
- 1 September 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Transportation Systems
- Vol. 23 (7) , 9747-9754
Click Fraud Detection of Online Advertising–LSH Based Tensor Recovery Mechanism
Online advertising utilizes the Internet technique to deliver marketing messages to promotional consumers. It is growing in recent years to facilitate the increasing demands of electronic commerce. Advertisers bid and pay for the advertisement whenever potential customers click it. The way of Pay-Per-Click (PPC) is vulnerable to malicious clicks that mimic real user behaviour to trick the platform into counting their clicks as legitimate. It causes massive financial losses on advertisers and also significantly reduces the credibility of online advertising platforms. The common strategies to detect fraud clicks are dynamically tailoring and interpreting data based on the machine learning model. These algorithms treat multi-dimensional data as an individual feature vector or matrix, making it different to explore intrinsic relations among a sequence of data. Million daily fraud clicks on various types further disperse the focus of models and result in relatively low efficiency for the current fraud prediction system. To tackle the fraud click problem, we introduce a tensor-based mechanism to predict fraud clicks. This paper considered reconstructing data into a high-rank tensor, implement tensor decomposition and transformation to explore hidden information under each data and explore the joined effect among a sequence of data. The proposed tensor transformation algorithm with locality-sensitive hashing (LSH) is tested by extensive experiments using real-world data. Compared with the state-of-art machine learning algorithms, our model can achieve significant performance in terms of accuracy and prediction-recall rate.
- 1 September 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Transportation Systems
- Vol. 23 (7) , 9747-9754
Click Fraud Detection of Online Advertising–LSH Based Tensor Recovery Mechanism
Online advertising utilizes the Internet technique to deliver marketing messages to promotional consumers. It is growing in recent years to facilitate the increasing demands of electronic commerce. Advertisers bid and pay for the advertisement whenever potential customers click it. The way of Pay-Per-Click (PPC) is vulnerable to malicious clicks that mimic real user behaviour to trick the platform into counting their clicks as legitimate. It causes massive financial losses on advertisers and also significantly reduces the credibility of online advertising platforms. The common strategies to detect fraud clicks are dynamically tailoring and interpreting data based on the machine learning model. These algorithms treat multi-dimensional data as an individual feature vector or matrix, making it different to explore intrinsic relations among a sequence of data. Million daily fraud clicks on various types further disperse the focus of models and result in relatively low efficiency for the current fraud prediction system. To tackle the fraud click problem, we introduce a tensor-based mechanism to predict fraud clicks. This paper considered reconstructing data into a high-rank tensor, implement tensor decomposition and transformation to explore hidden information under each data and explore the joined effect among a sequence of data. The proposed tensor transformation algorithm with locality-sensitive hashing (LSH) is tested by extensive experiments using real-world data. Compared with the state-of-art machine learning algorithms, our model can achieve significant performance in terms of accuracy and prediction-recall rate.
- 2 February 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 8 (7) , 5211-5218
Practical Network Coding Technologies and Softwarization in Wireless Networks
Network coding is an elegant and novel technique to improve network throughput and performance. It is considered as a critical technology to facilitate ever-increasing demands of future wireless networks. It exploits the broadcast nature of wireless media and cooperatively codes packets from different senders to provide reliable, secure, and efficient transmissions. Current research focuses on either transmission delay, coding complexity, forwarding security, or end-to-end throughput. Network coding-aided solutions can recover lost packets without feedback, eliminate latency, reduce the routing cost on diverse paths, or optimize the capacity of unstable wireless networks. However, devices or smart sensors usually have limited computational capacity and some applications could not tolerate high decoding delay, which prevents network coding from being widely deployed in the real world. In recent years, many research methods consider simplifying decoding matrix or coding algorithm to alleviate the shortcoming of network coding and further satisfy the extreme demands of the future wireless network. This article summarizes complexity-optimized methods and explains the interaction effect of coding opportunities and decoding overhead. We propose a taxonomy of practical network coding methods and illustrate three practical directions on cutting computational complexity and enhancing progressive decoding. We also conclude the benefit and cost of current network coding algorithms along with the outline of future research.
- 2 February 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 8 (7) , 5211-5218
Practical Network Coding Technologies and Softwarization in Wireless Networks
Network coding is an elegant and novel technique to improve network throughput and performance. It is considered as a critical technology to facilitate ever-increasing demands of future wireless networks. It exploits the broadcast nature of wireless media and cooperatively codes packets from different senders to provide reliable, secure, and efficient transmissions. Current research focuses on either transmission delay, coding complexity, forwarding security, or end-to-end throughput. Network coding-aided solutions can recover lost packets without feedback, eliminate latency, reduce the routing cost on diverse paths, or optimize the capacity of unstable wireless networks. However, devices or smart sensors usually have limited computational capacity and some applications could not tolerate high decoding delay, which prevents network coding from being widely deployed in the real world. In recent years, many research methods consider simplifying decoding matrix or coding algorithm to alleviate the shortcoming of network coding and further satisfy the extreme demands of the future wireless network. This article summarizes complexity-optimized methods and explains the interaction effect of coding opportunities and decoding overhead. We propose a taxonomy of practical network coding methods and illustrate three practical directions on cutting computational complexity and enhancing progressive decoding. We also conclude the benefit and cost of current network coding algorithms along with the outline of future research.
- 1 June 2020
- Published by Walter de Gruyter GmbH in Studies in Nonlinear Dynamics & Econometrics
- Vol. 25 (3) , 35-62
Learning for infinitely divisible GARCH models in option pricing
This paper studies the option valuation problem of non-Gaussian and asymmetric GARCH models from a state-space structure perspective. Assuming innovations following an infinitely divisible distribution, we apply different estimation methods including filtering and learning approaches. We then investigate the performance in pricing S&P 500 index short-term options after obtaining a proper change of measure. We find that the sequential Bayesian learning approach (SBLA) significantly and robustly decreases the option pricing errors. Our theoretical and empirical findings also suggest that, when stock returns are non-Gaussian distributed, their innovations under the risk-neutral measure may present more non-normality, exhibit higher volatility, and have a stronger leverage effect than under the physical measure.
- 1 June 2020
- Published by Walter de Gruyter GmbH in Studies in Nonlinear Dynamics & Econometrics
- Vol. 25 (3) , 35-62
Learning for infinitely divisible GARCH models in option pricing
This paper studies the option valuation problem of non-Gaussian and asymmetric GARCH models from a state-space structure perspective. Assuming innovations following an infinitely divisible distribution, we apply different estimation methods including filtering and learning approaches. We then investigate the performance in pricing S&P 500 index short-term options after obtaining a proper change of measure. We find that the sequential Bayesian learning approach (SBLA) significantly and robustly decreases the option pricing errors. Our theoretical and empirical findings also suggest that, when stock returns are non-Gaussian distributed, their innovations under the risk-neutral measure may present more non-normality, exhibit higher volatility, and have a stronger leverage effect than under the physical measure.