时间：2020年1月14日 上午9:30

地点:1A502

讲座题目：

Mixtures of Gaussian copula factor analyzers for clustering high dimensional data

摘要：

Mixtures of factor analyzers is a useful model-based clustering method which can avoid the curse of dimensionality in high-dimensional clustering. However, this approach is sensitive to both diverse non-normalities of marginal variables and outliers, which are commonly observed in multivariate experiments. We propose mixtures of Gaussian copula factor analyzers (MGCFA) for clustering high-dimensional clustering. This model has two advantages; (1)it allows different marginal distributions to facilitate fitting flexibility of the mixture model, (2)it can avoid the curse of dimensionality by embedding the factor analytic structure in the component-correlation matrices of the mixture distribution. An EM algorithm is developed for the fitting of MGCFA. The proposed method is free of the curse of dimensionality and allows any parametric marginal distribution which fits best to the data. It is applied to both synthetic data and a microarray gene expression data for clustering and shows its better performance over several existing methods.

Jangsun Baek教授个人简介：

Research Interests

• Bioinformatics

• Pattern Recognition

• Nonparametric Function Estimation

Education

• Ph.D., 1991, Statistics, Texas A&M University, U.S.A.

• M.S., 1984, Statistics, Yonsei University, South Korea

• B.A., 1981, Statistics, Yonsei University, South Korea

Award

• Connor Award, 1990, Dept. of Statistics, Texas A&M University, U.S.A