摘要
特征代码反病毒系统已经不能适应当前计算机病毒的检测要求,行为检测法将取代特征代码法。在对病毒恶意行为进行简要分析后,将状态跳转法应用到病毒的行为监测,提出基于概率训练的病毒检测模型。给出单个恶意行为和连续恶意行为病毒概率训练的具体方法,分析病毒概率阈值训练的过程,以此实现病毒判断算法。通过实例分析、检测结果验证基于概率训练的病毒检测模型的有效性。
Characteristic code anti-virus system can no longer meet current computer virus detection demand and will be replaced by the behavior detection method. With a brief analysis of virus abnormal behavior and an application of the state transition method to the virus behavior detection, this paper proposed the probability training based virus detection model. Implemented the virus judgment algorithm with a concrete virus probability training procedure of both single and continuous abnormal behavior, and a training procedure of virus probability threshold. The probability training based virus detection model proves to be effective with reliable case analysis and detection result verification.
出处
《计算机应用研究》
CSCD
北大核心
2010年第6期2316-2320,共5页
Application Research of Computers
基金
科技部国家"十一五"科技支撑计划项目(2007BAK34B06)
南京邮电大学攀登计划项目(NY208009)
关键词
病毒
检测模型
形式化
概率训练
virus
detection model
formalize
probability training