Abstract: This paper proposes a negative selection and danger-theory based assessment method for network risk is proposed. The danger signal is presented using cloud model. The antigen and antibody and their match are formulated. An improved backward cloud algorithm is used to produce the cloud numerical characteristics for assessment indicators of network risk. Experimental results prove that the proposed method can detect the intrusion attacks more effectively and reduce the alarm rate, then evaluate the network risk more credibly.

Key words: Artificial Immune System(AIS), danger signal, cloud model, network intrusion, risk assessment

摘要： 基于免疫否定选择和危险理论，提出一种网络入侵风险评估方法。采用云模型对危险信号进行描述，给出抗体、抗原的形式化定义和匹配过程，并利用一种改进的逆向云生成算法生成网络风险评估指标的云数字特征。实验结果表明，该方法可以更有效地检测网络攻击，降低虚警率，提高网络入侵风险评估的准确性。

关键词: 人工免疫系统, 危险信号, 云模型, 网络入侵, 风险评估

CLC Number: 

• TP18

• TP393