Kangying Zhou’s scientific contributions

What is this page?


This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.

Publications (4)


Expected out‐of‐sample R2$\bm {R^{2}}$ and norm of least‐squares coefficient. This figure shows the limiting out‐of‐sample R² and β̂$\hat{\beta}$ norm as a function of c and z from Proposition 3 assuming Ψ is the identity matrix and b∗=0.2$b_{*}=0.2$. [Color figure can be viewed at wileyonlinelibrary.com]
Expected out‐of‐sample risk and return of market timing. This figure shows the limiting out‐of‐sample expected return and volatility of the market timing strategy as a function of c and z from Proposition 3 assuming Ψ is the identity matrix and b∗=0.2$b_{*}=0.2$. [Color figure can be viewed at wileyonlinelibrary.com]
Expected out‐of‐sample Sharpe ratio of market timing. This figure shows the limiting out‐of‐sample Sharpe ratio of the market timing strategy as a function of c and z from Proposition 3 assuming Ψ is the identity matrix and b∗=0.2$b_{*}=0.2$. [Color figure can be viewed at wileyonlinelibrary.com]
Expected out‐of‐sample prediction accuracy from misspecified models. This figure shows the limiting out‐of‐sample R² and β̂$\hat{\beta}$ norm as a function of c and z from Proposition 6 assuming Ψ is the identity matrix, b∗=0.2$b_{*}=0.2$, and the complexity of the true model is c=10$c=10$. [Color figure can be viewed at wileyonlinelibrary.com]
Expected out‐of‐sample risk and return from misspecified models. This figure shows the limiting out‐of‐sample expected return and volatility of the market timing strategy as a function of c and z from Proposition 6 assuming Ψ is the identity matrix, b∗=0.2$b_{*}=0.2$, and the complexity of the true model is c=10$c=10$. [Color figure can be viewed at wileyonlinelibrary.com]

+6

The Virtue of Complexity in Return Prediction
  • Article
  • Full-text available

December 2023

·

42 Reads

·

67 Citations

Bryan Kelly

·

Semyon Malamud

·

Kangying Zhou

Much of the extant literature predicts market returns with “simple” models that use only a few parameters. Contrary to conventional wisdom, we theoretically prove that simple models severely understate return predictability compared to “complex” models in which the number of parameters exceeds the number of observations. We empirically document the virtue of complexity in U.S. equity market return prediction. Our findings establish the rationale for modeling expected returns through machine learning.

Download



Citations (4)


... Bianchi et al. (2021) state that the success of neural networks in bond price predictions is due to the model's ability to capture complex nonlinearities in the data. This is echoed by Kelly et al. (2024), who establish a theoretical underpinning to the observation that machine learning models, such as our neural network, outperform linear models. A single study to date by Kim et al. (2021) applies a battery of machine learning methods to predict corporate bond yield spreads. ...

Reference:

Does Real Estate Determine REIT Bond Risk Premia?
The Virtue of Complexity in Return Prediction

... The P-Tree framework has a notable extension, the random P-Forest, which connects to recent discussions on "benign overfitting" and "high-dimensional interpolation" (e.g., Belkin et al., 2019;Hastie et al., 2022) in statistics , as well as the corresponding "virtue of complexity" in financial contexts (e.g., Kelly et al., 2022Didisheim et al., 2024). We corroborate these studies by showing that large tree-based models for which the number of parameters exceeds the number of observations perform better OOS, provided that appropriate statistical regularization is applied. ...

The Virtue of Complexity Everywhere
  • Citing Article
  • January 2022

SSRN Electronic Journal

... Firstly, training data are often limited and the number of features that researchers can create is often much greater than the number of observations. In some research, such as [1], the ratio of the number of features over the number of observations, defined as model complexity can increase up to hundreds for financial instruments with a limited amount of history. Traditional setups in machine learning are not well-equipped for these data-scarce environments. ...

The Virtue of Complexity Everywhere
  • Citing Article
  • January 2022

SSRN Electronic Journal

... In the most recent literature it is possible to find studies related to the Implementation and Evaluation of Health Services (Møgster et al., 2024;Mbwasi et al., 2024;Davis et al., 2024;Naess et al., 2023), Financial Management and Performance Measurement (Tenzer, 2024;Kelly et al., 2023;Berlin et al., 2023), Governance and Policy Analysis (Huhtanen, 2024;Simonet, 2015;Eriksson and Andersson, 2023;Mori, 2023;Josewski, 2023), Health Workforce and Professional Roles (Byrne, 2024;Brenne, 2024;Maehreet al., 2024), Technological and Analytical Approaches (Nguyen, 2024;Al-Osaimi, 2024), Organizational Dynamics and Responses to NPM (van der Steen et al., 2024). ...

The Virtue of Complexity in Return Prediction
  • Citing Article
  • January 2022

SSRN Electronic Journal