Detection of MMORPG Misconducts Based on Action Frequencies, Types and Time-Intervals
Ruck Thawonmas, Yoshitaka Kashifuji · DMIN · 2010
This paper describes an application of data mining to detecting misconducts, such as use of bots and cheats, in Massively Multiplayer Online Role-Playing Games (MMORPGs). A method is proposed that exploits the information on action frequencies, types, and intervals available in the targeted MMORPG log data. The aforementioned information is used as the input to a support vector machine. Evaluation results, using log data, available in rollback database, from Cabal Online, confirm the effectiveness of the proposed method.