Exploring Abnormal Behavior Patterns of Online Users With Emotional Eating Behavior: Topic Modeling Study.

This study introduces a data-driven approach for analyzing behavior patterns of social website users with EE behaviors. We discovered the possibility of the LDA topic model as an exploratory user study method for abnormal behaviors in medical research. We also investigated the possibilities of ML- and topic modeling-based classifiers to automatically categorize text-based behavioral data, which could be applied to personalized medicine in future research.