基于风险反馈的网络安全密钥可信度识别仿真

Simulation of Network Security Key Credibility Identification Based on Risk Feedback

已有的网络密钥可信度识别方法无法准确设置风险反馈阈值,导致识别过程稳定性和误差均较差。为更精准找出网络安全密钥脆弱点,确保用户网络数据机密性与完整性,提出新的基于风险反馈的网络安全密钥可信度识别方法。使用椭圆曲线公钥密码算法创建网络安全密钥,从不同角度分析影响密钥风险的因素条件,利用XGBoost方法构建风险反馈模型。利用设置的风险反馈阈值明确密钥是否存在安全隐患;将网络安全密钥数据包变换成结构统一的格式,通过激活函数得到数据特征图,运用Mini-Batch随机梯度下降法划分密钥数据集,完成密钥可信度的识别。仿真结果证明,所提方法时间跨度为4时具备最优的可信识别精度,在恶意攻击下也能完成高密钥可信识别,实用性较强。

The existing methods are unable to accurately set the risk feedback threshold,resulting in poor stability and error during identification.In order to find out the vulnerability of network security keys more accurately and ensure the confidentiality and integrity of network data,based on risk feedback,this paper puts forward a new method to identify network security key credibility.First of all,the elliptic curve public key cryptography algorithm was used to create network security keys.And then,the risk factors and conditions affecting the key were analyzed from a different angle.Moreover,a risk feedback model was constructed by the XGBoost method.Furthermore,the risk feedback threshold was adopted to judge whether the key had potential risks.After that,the network security key packet was transformed into a unified format,and the data feature map was obtained through activation functions.Finally,the key data sets were divided by Mini-Batch stochastic gradient descent.Thus,we completed the identification of key credibility.Simulation results prove that the proposed method has the best credibility identification when the time span is four and can also complete the identification with high key creditability under malicious attacks,with strong practicability.