Leveraging Sociological Models for Predictive Analytics

TL;DR

This paper demonstrates that integrating sociological models into machine learning algorithms significantly enhances the prediction of human behavior and social dynamics, especially in adversarial contexts with limited data.

Contribution

It introduces sociologically-grounded learning algorithms that outperform standard methods in predicting relationships, viral diffusion, and future actions in adversarial social networks.

Findings

Sociological models improve prediction accuracy in social networks.

Algorithms perform well with limited training data.

Enhanced prediction of viral diffusion and adversarial actions.

Abstract

There is considerable interest in developing techniques for predicting human behavior, for instance to enable emerging contentious situations to be forecast or the nature of ongoing but hidden activities to be inferred. A promising approach to this problem is to identify and collect appropriate empirical data and then apply machine learning methods to these data to generate the predictions. This paper shows the performance of such learning algorithms often can be improved substantially by leveraging sociological models in their development and implementation. In particular, we demonstrate that sociologically-grounded learning algorithms outperform gold-standard methods in three important and challenging tasks: 1.) inferring the (unobserved) nature of relationships in adversarial social networks, 2.) predicting whether nascent social diffusion events will go viral, and 3.) anticipating and defending future actions of opponents in adversarial settings. Significantly, the new algorithms perform well even when there is limited data available for their training and execution.