The Journal JOE has achieved a 2025 Impact Factor of 21.7, and Top Tier journal in Econometrics. Doctoral students from the school of statistics and data science Zhang Tong has published the paper, entitled ”Estimation and Inference for Large-Dimensional Generalized Matrix Factor Models”, in this prestigious journal. The paper lists our university as the first affiliation, with Professor Kong Xinbing serving as the supervising author. The abstract and link to the full paper are provided below.
Estimation and Inference for Large-Dimensional Generalized Matrix Factor Models
Xinbing Kong;Tong Zhang
Abstract:This article introduces a nonlinear generalized matrix factor model, moving beyond the linear-Gaussian framework to accommodate a broader class of response models typically handled via logit, probit, Poisson, or Tobit structures. We introduce a novel Lagrange multiplier method carefully tailored to ensure that the penalized likelihood function is locally concave around the true factor and loading parameters. This leads to central limit theorems of the estimated factors and loadings which is nontrivial for nonlinear matrix factor modeling. We establish the convergence rates of the estimated factor and loading matrices for the generalized matrix factor model under general conditions that allow for correlations across samples, rows, and columns. We provide a model selection criterion to determine the numbers of row and column factors. Extensive simulation studies demonstrate the superiority in handling discrete and mixed-type variables of the generalized matrix factor model. An empirical data analysis of the company’s operating performance shows that the generalized matrix factor model does clustering and reconstruction well in the presence of discontinuous entries in the data matrix.
Link: https://www.sciencedirect.com/science/article/abs/pii/S0304407625002325
