Detecting insider trading of financial instruments is described. In an embodiment news event data is received from a sentiment index where each news event has at least one associated rating of a relationship between the news event and the financial instrument. For example, a numerical rating of confidence that the news event will influence volatility of the financial instrument and a numerical rating of how related a news event is to a particular company. In examples the news events are filtered using the ratings and remaining news events are used to identify potential insider trading where a price of an order within a time period before the news event differs by more than a threshold amount from a characteristic of the market after the news event. Parameters such as the time period and the threshold may be automatically adjusted using feedback data or may be dynamically user adjusted.
展开▼