PWPengfei Wang

Pengfei Wang

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Pengfei Wang has a total of 15 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Robust learning for ridge-penalized quasi-GLMs under non-identical distributions

    by Huiming Zhang, Wan Tian, Qiuran Yao, Pengfei Wang, Baochang Zhang

    In recent years, robust estimation of model parameters has attracted considerable attention in statistics and machine learning, particularly in the context of modeling data with outliers and related inverse problems. This paper introduces a novel log-truncated minimization estimator and a corresponding stochastic gradient descent (SGD) algorithm for quasi-generalized linear models. This approach provides a robust alternative to ordinary GLMs without assuming light-tailed error distributions. For independent non-identical distributed (i.n.i.d.) data, we derive non-asymptotic excess risk and \(\ell _2\)-risk bounds using log-truncated Lipschitz losses, assuming only a finite \(\beta\)-th moment for \(\beta \in (1,2]\). Notably, our estimator does not require higher moments or finite variance. Our contribution is to robustify the objective while allowing i.n.i.d. sampling. We also analyze the iteration complexity of SGD for finding stationary points of non-convex log-truncated minimization. Empirically, our SGD algorithm outperforms non-robust methods. We demonstrate its practical effectiveness through a real-data analysis of the German health care demand dataset using robust negative binomial regression.

  • Open Access

    Multiple Testing in Genome-Wide Association Studies via Hierarchical Hidden Markov Models

    by Pengfei Wang, Zhaofeng Tian

    The problems of large-scale multiple testing are often encountered in modern scientific researches. Conventional multiple testing procedures usually suffer considerable loss of testing efficiency due to the lack of consideration of correlations among tests. In fact, the appropriate use of correlation information not only enhances the efficacy of multiple testing but also improves the interpretability of the results. Since the disease- or trait-related single nucleotide polymorphisms (SNPs) often tend to be clustered and exhibit serial correlations, the hidden Markov model (HMM) based multiple testing procedure has been successfully applied in genome-wide association studies (GWAS). It is important to note that modeling the entire chromosome using one HMM is somewhat rough. To overcome this issue, this paper employs the hierarchical hidden Markov model (HHMM) to describe local correlations among tests and develops a multiple testing procedure that can not only automatically divide different class of chromosome regions, but also takes into account local correlations among tests. Theoretically, it is shown that the proposed multiple testing procedure is valid and optimal in some sense. Then a data-driven procedure is developed to mimic the oracle version. Extensive simulations and the real data analysis show that the novel multiple testing procedure outperforms its competitors.

  • Large‐scale covariate‐assisted two‐sample inference under dependence

    by Pengfei Wang, Wensheng Zhu

    The problems of large‐scale two‐sample inference often arise from the statistical analysis of “high throughput" data. Conventional multiple testing procedures usually suffer from loss of testing efficiency when conducting two‐sample ‐tests directly. To some extent, this is because of the ignorance of sparsity information. Moreover, the two‐sample tests commonly have local correlations, and neglecting the dependence structure may decrease the statistical accuracy. Therefore, it is imperative to develop a procedure that considers both sparsity information and dependence structure among the tests. We start by introducing a novel dependence model to allow for sparsity information and dependence structure. Based on the dependence model, we propose a covariate‐assisted local index of significance procedure and show that it is valid and optimal. Then a data‐driven procedure is developed to mimic the oracle procedure. Both simulations and real data analysis show that theCOALISprocedure outperforms its competitors.


  • Cauchy’s residue theorem for a class of real integrals

    by Zheng Tan, Pengfei Wang, Xiaoying Wang

    Residue theorem is a powerful tool in complex analysis to compute difficult real definite integrals. In this paper, with the help of special auxiliary functions, we’re going to use residue theorem to transform the calculation of real integrals into the calculation of residue. Moreover, we obtain a rather universal approach to solve this kind of problem. Finally, we apply this approach to reckon some specific infinite integrals and derive a special condition for a class of real integrals to be zero.

  • Large-scale multiple testing via multivariate hidden Markov models

    by Zhiqiang Hou, Pengfei Wang

    Large-scale multiple testing with correlated tests and auxiliary statistics arises in a wide range of scientific fields. Conventional multiple testing procedures largely ignored auxiliary information, such as sparsity information, and the dependence structure among tests. This may result in loss of testing efficiency. In this paper, we propose a procedure, called multivariate local index of significance (mvLIS) procedure, for large-scale multiple testing. The mvLIS procedure can not only characterize local correlations among tests via a Markov chain but also incorporates auxiliary information via multivariate statistics. We present that the oracle mvLIS procedure is valid, namely, it controls false discovery rate (FDR) at the pre-specified level, and show that it yields the smallest false non-discovery rate (FNR) at the same FDR level. Then a data-driven mvLIS procedure is developed to mimic the oracle procedure. Comprehensive simulation studies and a real data analysis of schizophrenia (SCZ) data are performed to illustrate the superior performance of the mvLIS procedure. Moreover, as a byproduct that is of independent interest, we generalize the single-index modulated (SIM) multiple testing procedure, which embeds prior information via 2-dimensional p-values, to allow for d-dimensional ( d≥3) statistics in multiple testing. The detailed extension is deferred to Discussion.