Tracking and tracing of goods movement is a key requirement for supply chain management and analysis. Data collection can be broad and large in volumes. Goods can moves in complex supply chain distributions, where disputes, frauds and thefts can happens. This paper aimed to develop a practical method to analyze the incoming data and employ unsupervised potential fraud detection in near real-time. The method is designed and discussed around peer group analysis (PGA) approach which is commonly used in financial market. The paper shall focus on two steps. First, monitor and groups good movements and categorize vendors or suppliers with similar trend / behaviours into dedicatedpeers. Second build a tool / services that detect anomalies in event transactions. The monitoring serviceshalldetect the outlier orindividual objects that distinct from peers whichpotentially fraud /alerts.
Copyrights © 2014