A0381
Title: Learning human activity patterns using clustered point processes with active and inactive states
Authors: Biao Cai - City University of Hong Kong (United States) [presenting]
Abstract: Modelling event patterns is a central task in a wide range of disciplines. In applications such as studying human activity patterns, events often arrive clustered with sporadic and long periods of inactivity. Such heterogeneity in event patterns poses challenges for existing point process models. A new class of clustered point processes is proposed that alternate between active and inactive states. The proposed model is flexible, highly interpretable, and can provide useful insights into event patterns. A composite likelihood approach and a composite EM estimation procedure are developed for efficient and numerically stable parameter estimation. Both the computational and statistical properties of the estimator, including convergence, consistency, and asymptotic normality, are studied. The proposed method is applied to Donald Trumps Twitter data to investigate if and how his behaviours evolved before, during, and after the presidential campaign. Additionally, large-scale social media data is analyzed from Sina Weibo and identify interesting groups of users with distinct behaviours.