Reconstructing online behaviors by effort minimization

  • Authors:

  • Armin Ashouri Rad;Hazhir Rahmandad

  • Affiliations:

  • Grado Department of Industrial and Systems Engineering, Virginia Tech, Northern Virginia Center, Falls Church, VA;Grado Department of Industrial and Systems Engineering, Virginia Tech, Northern Virginia Center, Falls Church, VA

  • Venue:
  • SBP'13 Proceedings of the 6th international conference on Social Computing, Behavioral-Cultural Modeling and Prediction
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

Much social interaction is moving online, offering new opportunities to analyze and understand fundamental patterns of human social behavior. One of the challenges in using this data is lack of direct observations of users' online activity in typical datasets. Building on the idea that people conserve their efforts in their online behavior, we develop a generic procedure for inferring user online behavior from their observable interactions with online objects and apply it to data from a social news website. We estimate which pages the users have seen and what stories they have observed. We test the effectiveness of this method in increasing the accuracy of a regression model that attempts to predict the number of votes a story is expected to receive, and show that the method can significantly increase the precision of these regressions.