According to the KRX, the key features of the upgraded system will include an improved unfair trading detection model; automation of analysis of the accounts suspicious of being involved in unfair trading, and adoption of various analytical tools such as the visual analysis system, which the surveillance officers can utilize in identification of abnormal trading patterns.
KRX said that the upgraded surveillance system took into account the characteristics of time series analysis and enhanced the reliability of statistical model by eliminating unnecessary over-detection of abnormal trading during certain time of the day.
To detecting an unfair trading practice, the criteria were subdivided based on the trading pattern and a score was assigned according to the degree of unfair trading practice per detection criterion.
If the accumulated score reaches to a predetermined level, the concerned trade or account will be identified as being associated with suspicious trading.
To detect the new types of unfair trading practices, both the rule-based model and statistical model were enhanced by introducing the tools based on artificial intelligence.
The newly upgraded market surveillance system automates the analysis of accounts suspicious of involved in or associated with unfair trading and makes possible the visual analysis of quotes and matching processes.
By developing the general screens for analysis, which are used by market surveillance officers, the accuracy of identifying unfair trading practices has been improved and the time spent for analysis is significantly reduced.
