The Computational Landscape of User Behavior on Social Media
Date & Time
Apr 13, 2017
from 02:15 PM to 03:05 PM
Kerr Hall 178
The rise of human interaction in digital environments has lead to an abundance of behavioral traces. These traces allow for model-based investigation of human-human and human-machine interaction ‘in the wild.’ Stochastic models allow us to both predict and understand human behavior. Dr. Darmon studies statistical procedures for learning such models from the behavioral traces left in digital environments.
Among other things he investigates how well user behavior can be attributed to time of day, self-memory, and social inputs. The models allow us to describe how a user processes their past behavior and their social inputs. He finds that despite the diversity of observed user behavior, most models inferred fall into a small subclass of all possible finitary processes. While online behavior of some users are quite complex (requiring models with almost 100 states to describe them), the behavior of 50% of users can be optimally described with 1-2 state hidden Markov models, some 60-70 % with simple renewal processes, and 95 % with models of up to 13 predictive states. This demonstrates that user behavior, while quite complex, belies simple underlying computational structures and therefore illustrates the potential for better understanding seemingly complex social information dynamics using only a limited number of simple models.