Design For Controlled Experiments For Social Networks With Covariates
Speaker:
Lulu Kang, Associate Professor of Applied Mathematics, Illinois Tech
Description:
A/B testing refers to the statistical procedure of conducting an experiment to compare two treatments applied to different testing subjects. The subjects participating in the online A/B testing experiments are users who are connected in social networks. Two connected users are similar in terms of their social behaviors. Hence, it is natural to assume that their reactions to online products and services are related to their network adjacency. We propose to use the conditional auto-regressive model to present the network structure and include the network effects and covariate effects in the estimation and inference of the treatment effect, and develop a D-optimal design accordingly.
Event Topic
Nonlinear Algebra and Statistics (NLASTATS)