Academic Report of the School of Mathematical Sciences [2025] No. 092
(High-Level University Construction Series Report No. 1114)
Lecture Title: Exponential Power Mixture of Experts Model: Estimation, Clustering, and Variable Selection
Lecturer: Associate Professor Feng Zhenghui (Harbin Institute of Technology, Shenzhen)
Lecture Time: October 11, 2025, 10:30-11:30 (morning)
Venue: Conference Room 307, Alumni Plaza
Abstract:
The mixture of experts model is a popular framework for modeling heterogeneity in data for regression, classification, and clustering. For regression and cluster analyses of continuous data, MoE usually uses normal experts following the Gaussian distribution. However, for a set of data containing a group or groups of observations with heavy tails or outliers, the use of normal experts is unsuitable and can unduly affect the fit of the MoE model. Motivated by the flexibility of the exponential power distribution, we propose EPMoE model and provide the corresponding estimation method, taking into account variable selection and the theoretical properties of the model. Numerical simulations demonstrate the effectiveness of our proposed method and algorithm, and finally, we provide a real data analysis to illustrate the application of our approach.
Lecturer’s Biography:
Feng Zhenghui, female, holds a Ph.D. in Statistics. She is an Associate Professor at the School of Science, Harbin Institute of Technology, Shenzhen, and serves as Director of the Professional Committee on Multivariate Analysis Applications of the Chinese Association of Applied Statistics. Her main research directions include functional data analysis, density estimation, big data, sufficient dimension reduction, and variable selection. She has published over 20 papers in international statistical journals such as Journal of the American Statistical Association and Computational Statistics and Data Analysis. She has presided over 4 national and provincial research projects, as well as 3 teaching projects. She once won the Excellent Teacher Award of the "Lianyi Award" in National Economic Management Experimental Teaching (Top 10 in China) and the First Prize in the 2nd Young Teachers' Teaching Competition of Harbin Institute of Technology, Shenzhen.
Faculty and students are welcome to attend!
Inviter: Wang Jiangzhou
School of Mathematical Sciences
September 30, 2025