Case-based reasoning for market surveillance

Stock market surveillance is an enormously difficult task. Exchange surveillance departments must sift through data on thousands of transactions each day, pick out suspicious price movements, investigate the transactions, and determine if action is necessary. Knowledge-based systems have been applied in this domain with mixed results, primarily due to a lack of clear-cut criteria for identifying anomalous activity. The authors present a novel approach to knowledge-based market monitoring: intelligent market monitor. The IMM is an application of case-based reasoning that evaluates trading patterns in the context of additional company, industry, and market data, and identifies possibly suspicious trades.<
>