AN ONLINE GAME ACCOUNT PROTECTION MODEL BASED ON FUZZY RECOGNITION OF USER BEHAVIORS

摘要

In recent years players have spent a lot of time and money on the online games, and the game accounts have become an important type of network virtual asset. However, due to the lack of effective security-mechanism, the current account protection is mainly-based on password verification. Once an user's password is leaked, its game assets will be illegally stolen and difficult to recover. Aiming at the problems, this paper proposed an account protection model based on fuzzy recognition of players' game behaviors. We built a fuzzy decision tree of the player behavior characteristics to identify illegal account users, while allowing normal user behavior to fluctuate within a reasonable range. Furthermore, considering the natural improvement of the player's skills, we introduced a periodic feedback-update mechanism to regularly renew the fuzzy decision tree with the newly collected behavior data, which realized dynamically learning of the users' behaviors. The experimental results based on the data collected from real games show that the proposed method can detect the abnormal user behaviors in a dynamical scenarios and protect the game assets from being illegally used.