基于改进深度学习的网络敏感信息快速过滤研究

Research on fast filtering method of network sensitive information based on improved deep learning

为了提高网络敏感信息的检测和过滤能力,提出基于改进深度学习的网络敏感信息快速过滤方法.构建网络敏感信息的数据存储结构模型,利用量化参数分析以及显著性检测的方法进行网络敏感信息采集,然后通过多尺度特征分解方法实现对网络敏感信息的信息融合,结合自相关融合聚类分析提取网络敏感信息的边缘分布特征量,实现网络敏感信息特征检测优化.在此基础上提取网络敏感信息数据的包络参数,通过改进深度学习方法实现对网络敏感信息的快速检测和快速过滤.仿真结果表明,采用该方法进行网络敏感信息过滤的可靠性较高,过滤收敛性水平较高,提高了网络敏感信息过滤能力.

In order to improve the detection and identification of sensitive information in overlapping community networks,it is necessary to filter the sensitive information on the network.A fast filtering method based on improved deep learning is proposed.Quantitative parameter analysis and saliency detection methods are used to construct a data storage structure model for sensitive information of overlapping community networks,and to realize the collection of sensitive information of overlapping community networks.Overlapping community network sensitive information feature detection optimization is realized through multi-scale feature decomposition method for information fusion of network sensitive information and autocorrelation fusion clustering analysis is used to extract edge distribution feature quantity of overlapping community network sensitive information.The envelope parameters of the sensitive information data of the overlapping community network are extracted and the improved deep learning method is adopted to realize the rapid detection and information fusion filtering of the sensitive information of the overlapping community network.The simulation results show that the method is used to filter the sensitive information of the overlapping community network with high reliability,and the filtering convergence level is better,and the detection ability of the sensitive information of the overlapping community network is improved.