Abstract Details
Activity Number:
|
301
|
Type:
|
Contributed
|
Date/Time:
|
Tuesday, August 5, 2014 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Section on Statistical Learning and Data Mining
|
Abstract #313715
|
|
Title:
|
WITHDRAWN: Real-Time Novelty Effect Detection for A/B Testing
|
Author(s):
|
Tianhong He and Yi Liu and Luo Lu
|
Companies:
|
Twitter and Twitter and Twitter
|
Keywords:
|
A/B testing ;
Novelty Effect ;
Experimental Design ;
Social Networks
|
Abstract:
|
A/B testing is a widely used practice for social networks with large audiences - a small fraction of users will be exposed to a new feature and response is measured to evaluate the effect. However, existence of confounding variables that correlate with both the dependent variable and the independent variable could lead to misestimation of experiment results. For example, novelty effect is one type of such confounding factors.
In this work, we develop a way to address this problem. First, we describe a novel method that tracks backwards through the whole experiment period to detect and eliminate novelty impact. We will show the advantages and limitations of this method, especially under the scenario of social networks. Second, by proposing to use event time instead of real time, we conquer the limitations in previous methods and further extend our algorithm to make real time novelty impact detection possible.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2014 program
|
2014 JSM Online Program Home
For information, contact jsm@amstat.org or phone (888) 231-3473.
If you have questions about the Professional Development program, please contact the Education Department.
The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.
Copyright © American Statistical Association.