An expert real-time trading system.

摘要

This thesis explores the use of a rule-based expert system with real-time market indexes and stock quotes as input, to predict trends in order to maximize gains and minimize losses. The expert system developed here uses a set of forward-chaining rules based on the comparison of past behavior and current real-time market data, coupled with the use of the Relative Strength Index to derive decisions for the purpose of day-trading. It also features a set of parameters that can be easily adjusted by the user to change the behavior of the rule set. Tests using daily data for the month of April 1998 are analyzed. Performance is also measured against the Dow Jones Industrial Average and the NASDAQ Composite Index. A suggestion for a centralized Internet stock trading service is briefly discussed.