Marking the Close analysis in Thai Bond Market Surveillance using association rules

This study investigates the opportunity of employing data mining techniques as a supplement to traditional techniques, such as economic modeling, to detect misconduct in the Thai Bond Market. In the study, association rules are used to detect “Mark the Close” conduct in the Thai Bond Market Association (ThaiBMA) surveillance system. The experiment was conducted on 54,334 trading transactions reported to the ThaiBMA in the year 2005. The result shows that association rules can be effectively used to provide a short list of traders who are likely to behave in such a way.