基于BP神经网络在物联网多敏感信息同源检索

Homologous Retrieval of Multi Sensitive Information in Internet of Things Based on BP Neural Network

针对物联网中多用户相似源敏感信息检索方法运算速度慢、结果质量较差问题,提出一种物联网多敏感信息同源检索算法。准确分析物联网各应用层,构建隐层的3层BP神经网络模型,引入动态因子的学习算法,结合多维仿生信息理论的点同源连续性规则,推导出多敏感信息同源相似度特征,排序信息同源相似度特征的相关性级别并输出,最终完成物联网多敏感信息同源检索。实验结果表明:所提方法运算时间较短,能够实现高效的多敏感信息同源检索,且用户对检索结果的满意度较高,提高了物联网多敏感信息同源检索的结果质量。

In order to solve the problems of slow computing speed and poor result quality of multi-user similar source sensitive information retrieval method in the Internet of things,this paper proposed a homologous retrieval algorithm of multi-sensitive information in the Internet of things.Firstly,we accurately analyzed each application layer of the Internet of things,and built a three-layer BP neural network model.Secondly,we introduced the learning algorithm based on dynamic factors and combined with the continuity rule of point homology of multi-dimensional bionic information theory to deduce the homologous similarity feature of multi-sensitive information.Thirdly,we sorted the correlation levels of homologous similarity feature of information and output them.Finally,we completed the homology retrieval of multi-sensitive information of the Internet of things.Simulation results prove that the proposed method has shorter operation time,higher efficiency of homology retrieval of multi sensitive information.In addition,the user has high satisfaction of retrieval result,and the quality of multi-sensitive information retrieval is improved.