Rolling Window Selection for Three-Pass Regression Filter Forecasting with Time-varying Coefficients (with Qiankun Zhou), accepted at Journal of Financial Econometrics [Latest Paper]
In this paper, we introduce smooth time-varying coefficients to the three-pass regression filter (3PRF) forecasting method proposed by Kelly and Pruitt (2015). Following Inoue et al. (2017), we use a rolling window selection to estimate parameters and generate forecasts. This approach uses only the most recent observations, addressing the trade-off between forecast bias and variance. We establish an optimal rate for selecting recent observations. Monte Carlo simulations show that, in general, the rolling window selection method for 3PRF forecasting with time-varying coefficients yields smaller forecasting mean square errors compared to both the original 3PRF method and the full-sample approach. Empirical applications to forecasting eight macroeconomic variables in the U.S. and the market return of the U.S. stock market underscore the practical advantages of our proposed approach.
Time-varying Factor-augmented Regression Model for High-dimensional Data [Latest Paper]
This paper investigates advancements in the factor-augmented regression model by incorporating time-varying factor loadings and forecasting coefficients. In the proposed model, factors are extracted from a large number of predictors using the boundary-corrected kernel method. Subsequently, these estimated factors are integrated into the time-varying forecast using the locally constant nonparametric method. An essential contribution of this study is the verification of the asymptotic consistency of the constructed forecasting target variable. Monte Carlo simulations corroborate the enhanced forecasting performance. Furthermore, empirical applications are conducted to forecast U.S. macroeconomic variables. The empirical findings underscore the reliability and precision of the proposed time-varying factor-augmented regression model.
Stochastic Learning of Semiparametric Monotone Index Models with Large Sample Size [Latest Paper]
This paper examines the estimation of semiparametric monotone index models when the sample size $n$ is extremely large and conventional approaches are inapplicable due to substantial computational burdens. Motivated by the Mini-Batch Gradient Descent algorithm (MBGD), this paper proposes a novel subsample- and iteration-based estimation procedure. In particular, starting from any initial guess of the true parameter, the estimator is progressively updated using a sequence of subsamples randomly drawn from the data set whose sample size is smaller than $n$. The update is based on the gradient of some well-chosen loss function, where the nonparametric component in the model is replaced with its Nadaraya-Watson kernel estimator. The key novelty of the new algorithm is that the kernel estimator is also constructed using random subsamples so that the algorithm is fully subsample-based. Compared with the full-sample-based iterative method, the new method reduces the computational time by roughly $n$ times if the subsample size and the kernel function are chosen properly, so can be easily applied when the sample size $n$ is large. The new estimator substantially improves the computational speed while at the same time maintains the estimation accuracy.
Urban Fire Risk Evaluation Integrating Image Features and Interpretable Machine Learning Models (with Zherui Li, Juncheng Jiang, and Wen Chen, Applied Spatial Analysis and Policy 18, no. 4 (2025): 151.
Pledge to Green: The Green Innovation Effect of Intellectual Property Pledge Financing Pilot Policy (with Tie Shi and Junbing Xu) Applied Economics (2024):1-13. [link]
Exploring the Spatial Pattern and Influencing Factors of Intercity Capital Flows from 2005 to 2019: A Case Study of Yangtze River Delta Region, China (with Zherui Li and Feng Zhen) Growth and Change,55, no.1 (2024): e12694. [link]
Urban Internal Network Structure and Resilience Characteristics from The Perspective of Population Flow: A Case Study Based on Nanjing, China (with Zherui Li, Wen Chen, and Zhe Cui) International Journal of Geo-Information 13, no.9 (2024):331. [link]
Green to health: The impact of environmental regulation on health status (with Junbing Xu and Yuning Wang) Sustainable Cities and Society 98 (2023): 104839. [link]
The VIX’s Impact on China’s Stock Market (with Yang-chao Wang and Jui-Jung Tsai) Journal of Business & Economics 6, no. 1 (2014): 23 [link]