Behavior Identification in Markets - using Visualization and Network Analysis

Market manipulation is an open problem in prediction markets. In this work a new method is introduced to identify fraud via analysis of the network structure. Furthermore, an interactive graph visualizations technique for market monitoring is developed. The results show that a low false positive rate is possible to attain, while keeping the number of true positives high. The suggested indicators are fast enough for real-time monitoring on small and medium sized graphs.