The NASD Securities Observation, New Analysis and Regulation System (SONAR)

Book One

Volume

Proceedings of the AAAI Conference on Innovative Applications of Artificial Intelligence Conference, 15

Deployed Applications

The Securities Observation, News Analysis, and Regulation (SONAR) system was developed by NASD to monitor the Nasdaq, Over the Counter (OTC), and Nasdaq-Liffe (futures) stock markets for potential insider trading and fraud through misrepresentation. SONAR has been in operational use at NASD since December 2001, processing approximately 10,000 news wires stories and SEC filings, evaluating price/volume models for 25,000 securities, and generating 50-60 alerts (or "breaks") per day for review by several groups of regulatory analysts and investigators. In addition, SONAR has greatly expanded surveillance coverage to new areas of the market and increased accuracy significantly over an earlier break detection system. SONAR makes use of several AI and statistical techniques, including NLP text mining, statistical regression, rule-based inference, uncertainty, and fuzzy matching. Sonar combines these enabling technologies in a system designed to deliver a steady stream of high-quality breaks to the analysts for further investigation. Additional components including visualization, text search agents, and flexible displays add to the system’s utility. Finally, SONAR is designed as a knowledge-based system. Domain knowledge is maintained by knowledge engineers working closely with the regulatory analysts. In this way, SONAR is adaptable to new market conditions, data sources, and regulatory concerns.