关于异构型物联网重复数据有效清除仿真  

Simulation of Effective Removal of Duplicate Data in Heterogeneous Internet of Things

作　　者：严圣华 王昌达[1] YAN Sheng-hua,WANG Chang-da(Graduate School, Jiangsu University, Jiangsu Zhenjiang 212016, Chin)

机构地区：[1]江苏大学研究生学院,江苏镇江212016

出　　处：《计算机仿真》2018年第5期444-447,共4页Computer Simulation

摘　　要：对异构型物联网的重复数据的清除,可以有效提高物联网运行稳定性。对异构物联网的重复数据进行有效清除,需要对全部的特征向量进行量化处理,进而组建字符关系矩阵,完成重复数据的有效清除。传统方法利用批处理块级重复数据清除方法计算出数据间的相似度,给出相似度判断阈值,但忽略了字符关系矩阵的建立,导致数据去重精度偏低。提出基于多维数据聚类的重复数据清除方法。将全部感知数据属性中的连续值进行离散化,提取各个数据文本中的特征向量,对全部的特征向量进行量化处理,融合傅立叶转换方法组建字符关系矩阵,给出每个字符和数据间的映射关系,计算每个数据的傅立叶系数向量,得到数据相似度判断阈值,由此实现面向异构型物联网的重复数据清除。仿真证明,所提方法清除精度高,有效地提升了异构型物联网环境下重复数据的质量。The aim of this research is to overcome defect of traditional elimination method for duplicated data of heterogeneous Internet of Things, such as poor elimination precision. Based on multi-dimensional data cluster, a new elimination method is proposed. Firstly, discretization is carried out for continuous value in all perception data attribute, and characteristic vector in each data text is extracted. Then, quantification process is carried out for all characteristic vectors Integrated with Fourier transform method, character relationship matrix is built and mapping relation between each character and data is provided. Fourier coefficient vector of each data is calculated to obtain judgment threshold of data similarity. Thus, the elimination of duplicated data is completed. Simulation proves that the method has high elimination precision. It improves quality of the duplicated data effectively.

关 键 词：异构型物联网 重复数据 有效清除 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]