Master's Defense

Recommendation Engine for Twitter Users

Speaker:Sriram Vepuri
sriram at cs.duke.edu
Date: Thursday, April 7, 2016
Time: 10:45am - 12:15pm
Location: North 306, Duke

Abstract

Currently, Twitter generates a major chunk of its revenue from ads. It places ads in the twitter feeds of users. To segment these users, it relies on interest graph gathering details from the tweets made by the users and also the sites these users visit. However, this approach directly doesn�t address the three key components of marketing i.e. right user, right time, and right ad. In this project, I propose to build an application, which will leverage the topic modeling aspects of machine learning to place a right ad or recommendation to the user.
Advisor(s): Rong Ge
Committee: Salman Azhar, Jesko Von Windheim