University of Connecticut

Events Calendar

Colloquium

Wednesday, October 22, 2014
4:00pm – 5:00pm

Storrs Campus
AUST 105

Professor Haiying Wang University of New Hampshire "Leveraging for Logistic Regression with Big Data" ----- For Big Data with large sample size n, it is computationally infeasible to obtain the maximum likelihood estimates for unknown parameters, especially when the maximum likelihood estimates does not have a close-form solution. This paper proposes random sub-sampling algorithms to efficiently approximate the maximum likelihood estimates of unknown parameters in a logistic regression model with binary responses, one of the most commonly used models in practice for classification. We theoretically prove the consistency of the algorithms, develop two optimal sub-sampling strategies and evaluate the performance of the proposed methods using synthetic and real data sets.

Contact:

Ellis Shaffer, ellis.shaffer@uconn.edu

Statistics (primary), UConn Master Calendar

Control Panel