|
Activity Number:
|
106
- New Frontiers and Developments in Causal Inference
|
|
Type:
|
Invited
|
|
Date/Time:
|
Monday, August 9, 2021 : 1:30 PM to 3:20 PM
|
|
Sponsor:
|
Section on Statistical Learning and Data Science
|
|
Abstract #317005
|
|
|
Title:
|
Small Weights for Big Data: Computational Aspects and Empirical Performance
|
|
Author(s):
|
Jose Zubizarreta* and Kwangho Kim
|
|
Companies:
|
Harvard University and Harvard University
|
|
Keywords:
|
|
|
Abstract:
|
We analyze computational aspects and illustrate the empirical performance of minimal dispersion approximately balancing ('small') weights in big data sets. We show how specialized alternating direction method of multipliers (ADMM) algorithms can be used with such weighting schemes to effectively adjust for covariates in large-scale observational studies, with millions of observations in minutes.
|
Authors who are presenting talks have a * after their name.
