On synthetic data generation for anomaly detection in complex social networks

TL;DR

This paper presents a novel agent-based modeling framework for generating synthetic social network data to improve anomaly detection, especially in scenarios with limited real data for mission-critical applications.

Contribution

It introduces a heuristic, social-knowledge-based ABM prototype for synthetic data generation tailored to anomaly detection in complex social networks.

Findings

The ABM prototype can generate data resembling real social network anomalies.

The model is useful for studying patterns of life and detecting anomalies with scarce data.

Directions for hyper-parameter tuning to improve data distribution similarity.

Abstract

This paper studies the feasibility of synthetic data generation for mission-critical applications. The emphasis is on synthetic data generation for anomalous detection in complex social networks. In particular, the development of a heuristic generative model, capable of creating data for anomalous rare activities in complex social networks is sought. To this end, lessons from social and political literature are applied to prototype a novel implementation of the Agent-based Modeling (ABM) framework, based on simple social interactions between agents, for synthetic data generation in the context of terrorist profile desegregation. The conclusion offers directions for further verification, fine-tuning, and proposes future directions of work for the ABM prototype, as a complex-societal approach to synthetic data generation, by identifying heuristic hyper-parameter tuning methodologies to further ensure the generated data distribution is similar to the true distribution of the original data-sets. While a rigorous mathematical optimization for reducing the distances in distributions is not offered in this work, we opine that this prototype of an autonomous-agent generative complex social model is useful for studying and researching on pattern of life and anomaly detection where there is strict limitation or lack of sufficient data for using practical machine learning solutions in mission-critical applications.