Factor Models for High-dimensional Count Data

Time：Fri., 9:30-10:30 am, Nov. 28, 2025

Venue：C654, Shuangqing Complex Building A

Organizer：Yunan Wu

Speaker：Tao Wang

Statistical Seminar

Organizer：

Yunan Wu 吴宇楠 (YMSC)

Speaker：

Tao Wang 王涛

上海交通大学

Time：

Fri., 9:30-10:30 am, Nov. 28, 2025

Venue：

C654, Shuangqing Complex Building A

Title:

Factor Models for High-dimensional Count Data

Abstract:

This talk presents recent advances in factor modeling for multivariate count data. We propose a maximum variational likelihood approach for estimation and inference under a multinomial factor-augmented inverse regression model, with asymptotic properties established in high-dimensional settings. For Poisson factor models, we introduce a data-driven criterion for determining the number of factors and prove its consistency as both sample size and dimensionality diverge. Extensions to zero-inflated models are also briefly discussed.

DATENovember 29, 2025

Related News

0
Factor Modeling for Clustering High-dimensional Time Series
AbstractWe propose a new unsupervised learning method for clustering a large number of time series based on a latent factor structure. Each cluster is characterized by its own cluster-specific factors in addition to some common factors which impact on all the time series concerned. Our setting also offers the flexibility that some time series may not belong to any clusters. The consistency with...
1
The form factor program: asymptotic factorization of n-particle SU(N) form factors
AbstractWe investigate the high energy behavior of the SU(N) chiral Gross-Neveu model in 1 + 1 dimensions. The model is integrable and matrix elements of several local operators (form factors) are known exactly. The form factors show rapidity space clustering, which means factorization, if a group of rapidities is shifted to infinity. We analyze this phenomenon for the SU(N) model. For several ...