Advances in Intelligence and Security Informatics

Chapter 5

Forecasting Complex Group Behavior via Multiple Plan Recognition

In recent years, group behavior forecasting has been attracting attention in a number of ISI research and application fields, such as homeland and national security, government policy evaluation, market analysis, pandemics and disaster response planning, to name just a few. Research on group behavior forecasting has mainly focused on building predictive models based on statistical learning methods [1, 2]. Although machine learning methods provide a computational means to construct predictive models, they heavily rely on the structured data, which are usually hard to obtain. Furthermore, because of their data-driven features, the explanation of prediction results given ...