基于邻域相似性的多源异构大数据属性约减方法

Attribute Reduction Method of Multi-source Heterogeneous Big Data Based on Neighborhood Similarity

针对网络多源异构大数据之间领域关系难以界定导致运算量较大的问题,提出基于邻域相似性的多源异构大数据属性约减方法。首先,基于多源异构大数据进行分析,根据边界域理念及邻域相似性提出属性重要度度量方法;其次,结合邻域粗糙集提出多源异构大数据的属性约减算法;最后,进行实验对比分析。实验结果表明,该方法可对多源异构大数据属性进行约减,且约减后的分类准确率更高,优于对比方法,具有良好的应用性能。

Aiming at the problem that the domain relationship between multi-source heterogeneous big data in the network is difficult to define,resulting in a large amount of computation,a method of attribute reduction of multi-source heterogeneous big data based on neighborhood similarity is proposed.Firstly,based on the analysis of multi-source heterogeneous big data,the attribute importance measurement method is proposed according to the concept o£boundary domain and neighborhood similarity.Secondly,the attribute reduction algorithm of multi-source heterogeneous big data is proposed based on the neighborhood rough set.Finally,the experiment is compared and analyzed.The experimental results show that this method can reduce the attributes of multi-source heterogeneous big data,and the reduced classification accuracy is higher,which is superior to the comparison method,and has good application performance.