大数据交叉映射融合的逆向云算法仿真  

Reverse Cloud Algorithm Simulation of Big Data Cross Mapping Fusion

作　　者：李杰 邹慧兰[2] LI Jie;ZOU Hui-lan(Institute of Computer and Remote Sensing Information Technology,North China Institute of Aerospace Engineering,Langfang Hebei 065000,China;School of Information Engineering,Minzu University of China,Beijing 100081,China)

机构地区：[1]北华航天工业学院计算机与遥感信息技术学院,河北廊坊065000 [2]中央民族大学信息工程学院,北京100081

出　　处：《计算机仿真》2020年第2期284-288,共5页Computer Simulation

基　　金：河北省青年科学基金项目(D2018409029);河北省廊坊市科技局资助项目(2018011002)。

摘　　要：大数据融合可以提升工作效率、保证数据安全。逆向云算法根据数据之间的关联性交叉映射,对不同类型数据执行逆向处理。而传统逆向云算法,分析大数据离散度的能力不强,致使计算结果的偏差较大,因此研究大数据交叉映射融合的逆向云算法。算法将原有算法作为基本理论,利用梯度联合函数反演,根据交叉梯度约束条件确定数据离散形式,捕捉数据交叉规则;依照已知节点数据的物理量,对未知节点数据插值,通过源数据项、目标数据项建立函数映射关系,得出隐含关联;按照关联度匹配数据特征,域隐含层特征函数对域网络参数反向微调,至此实现逆向云算法。实验结果可知:与传统的基于鲁棒性的大数据逆向云算法相比,上述算法分析大数据之间离散度的能力更强,得到的参数结果偏差极小。由此可见,所研究的逆向云算法的计算结果更加精确。Big data fusion can improve work efficiency and ensure data security.According to the associated cross mapping between data,the backward cloud algorithm performs the reverse processing on different types of data.Traditionally,the ability of algorithm was not strong for analyzing the dispersion of big data,resulting in large deviation of calculation results.Therefore,this article focuses on the backward cloud algorithm based on big data cross-mapping fusion.This algorithm took the original algorithm as the basic theory,and adopted the gradient joint function for inversion.And then,the discrete form of data and the crossing rule of data were determined by the cross gradient constraint conditions.According to the physical quantity of known node data,the data of unknown node was interpolated.Moreover,the function mapping relationship was established by source data items and target data items,and thus the implicit correlation was obtained.According to the correlation degree,the data features were matched.Finally,the domain network parameters were fine-tuned reversely by the feature function in domain hidden layer.Thus,the backward cloud algorithm was achieved.Simulation results show that,compared with traditional big data backward cloud algorithm based on robustness,the proposed algorithm has stronger ability to analyze the dispersion degree between big data.The deviation of parameter results is very small,so the calculation results of the proposed algorithm are more accurate.

关 键 词：大数据 交叉映射 数据融合 逆向云算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]