|
Activity Number:
|
444
- Recent Advances in Statistical Methodology for Big Data
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 12, 2021 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
IMS
|
|
Abstract #318387
|
|
|
Title:
|
An Efficient Monte Carlo EM Algorithm for Estimating Heterogenous Crossed Effects in GLMM
|
|
Author(s):
|
Rashmi Ranjan Bhuyan* and Gourab Mukherjee and Wreetabrata Kar
|
|
Companies:
|
University of Southern California and University of Southern California and Purdue University
|
|
Keywords:
|
Monte Carlo EM;
High dimensional;
Clustering;
Logistic regression;
Class Imbalance
|
|
Abstract:
|
The use of generalized linear mixed?effects models (GLMM) is popular across social and medical sciences. High dimensionality and class imbalance are two extremely common characteristics in such datasets. Under these settings, fitting GLMMs and estimating all the effects become immensely difficult. We propose a Monte Carlo EM algorithm to recover the heterogeneous effects in imbalanced GLMM. In addition to estimating the model, our algorithm also inherently implements clustering. We provide theoretical guarantees for consistent estimates of the effects. The algorithm is designed to work in the high dimensional regime, and it scales up efficiently. We compare the performance of the proposed algorithm on a range of simulation setups and observe encouraging performance. We demonstrate the applicability of our method by analyzing consumer behavior in email marketing dataset.
|
Authors who are presenting talks have a * after their name.
