Name classification in noisy domains

Many problems in trading and regulation depend on textual data that identifies objects, events, and relationships. In general, these trading and regulatory activities require data from a variety of sources that must be `intelligently integrated´, in order to draw attention to a particular object or group of objects to indicate unusual trading opportunities or to help indicate unusual market behavior for regulators. Text classification techniques have been used for a number of years on Wall Street as significant components of intelligent data integration systems. The authors discuss work on text classification for noisy domains, and discuss how this work has been applied in trading and regulatory systems