Research Colloquium 

Predicting Individual Behavior with Social Networks

• Daniel G. Goldstein, London Business School  & Yahoo Research
• Wednesday, November 3, 2010 2:00 to 4:00 pm
• Babbio 104

Predicting behavior is a fundamental objective of the social sciences. With the recent availability of social network data, it has become possible to predict an individual's behavior based on the activities of his or her social contacts. To advance the predictive abilities and knowledge of the sciences, however, such predictions must improve upon existing baseline models. I employ an online network of over 100 million people to predict a diverse range of individual-level behaviors from patronizing an offline department store to joining a recreational league.  In each instance, predictions of initial adoption were substantially improved by incorporating social data. I discuss factors that may govern the usefulness of social data for prediction, including whether initial or repeat behaviors are being predicted, and the nature of available predictors.

For more information, please contact:

Sharen Glennon
Associate Director, Marketing
Babbio Center
Room 4434
Phone: 201-216-5381
Fax: 201-216-5385
sglennon@stevens.edu

The Innovation University ^TM

Event flyer