Skip to Main Content
This paper describes a service architecture for a financial market monitoring and surveillance system in which different components interact in coordination with internal and external service providers to produce proactive alarms for potential fraud cases. The proposed service system is demonstrated through an exemplar case study of text mining and data mining to analyze the impact of 'stock-touting' spam e-mails and misleading press releases on trading data. It also shows how an independent service provider could have helped by raising the alarm about a potential on-going 'pump and dump' scheme. The proposed service architecture extends the Market Monitoring Framework (MMF) , by incorporating automated linguistics-based text mining techniques to extract the key concepts of spam e-mails and press releases, relate them with other available information and highlight any signs that they might be part of 'stock touting campaigns'. The analysis of emails and press releases through text mining components could help to raise proactive alarms that would not have been possible otherwise. Evaluation of the proposed service system is carried out through a case study that relates to a real case from the over-the-counter (OTC) market, and which was prosecuted by the SEC. Through this case, the paper goes on to explain how the proposed approach could be used within the existing fraud analysis process, the extent to which the process could be automated, the relationships with other types of analysis and the role that fraud analysts play.