复杂式网络用户隐私数据多层分类存储仿真

Multi-Layer Classification Storage Simulation of Complex Network User Privacy Data

为了确保复杂式网络用户隐私数据多层分类存储的完整性和有效性,针对当前方法数据存储的不完整和数据存储时间长的问题,提出基于贝叶斯的复杂式网络用户隐私数据多层分类存储方法。为了提升用户隐私数据多层分类存储的效果,通过对用户隐私数据权重的计算建立数据多层分类存储模型,在此基础上,分析用户隐私数据的特征向量和数据类别向量的联合分布情况,根据分布结果和概率分类器计算用户隐私数据的条件概率分布,并分析数据的对数似然概率,得到用户隐私的训练数据和测试数据,对数据类别展开预测,完成用户隐私数据的多层分类,利用数据文件的分块数、副本系数以及用户隐私数据最小的副本系数目标函数与数据分类结果相结合,最终实现用户隐私数据多层分类存储。实验结果表明,提出方法能够完整存储用户隐私数据,并且数据存储时间较短,降低了用户隐私数据多层分类存储能耗,验证了提出方法的有效性。

In order to ensure the integrity and validity of multi-layer classification storage of user privacy data in complex network,this article proposed a method of multi-layer classification storage for user privacy data in complex network.The multi-layer classification storage model of data was built by calculating the weight of user privacy data.On this basis,the joint distribution of feature vectors and data category vectors was analyzed.The distribution result and the probability classifier were used to calculate the conditional probability distribution of user private data and analyze the log-likelihood ratio of data,and then training data and test data of user privacy were obtained.Moreover,the data category was predicted,so as to complete the multi-layer classification of user private data.Finally,the multi-layer classification storage for the user privacy data was achieved by using the number of blocks of the data file,the back-up copy coefficient,and the objective function with the smallest copy coefficient of user privacy data and the data classification result.Simulation results show that the proposed method can completely store user privacy data,and the data storage time is shorter.Meanwhile,this method reduces the energy consumption of multi-class classification storage of user privacy data,so that the effectiveness of proposed method can be verified.