基于复杂网络数据流密度的增量子空间数据挖掘算法

Incremental subspace data mining method based on data stream density of complex network

为了提升在复杂网络中对大规模网络数据流进行挖掘时的准确性,提出一种基于复杂网络数据流密度的增量子空间数据挖掘算法,在算法中先对复杂网络的数据流密度进行分析,并根据不同网络的数据流密度来划分社区,进行无向环路遍历来确定数据流的所属社区。再通过增量子空间数据挖掘算法来计算社区网络与数据流的相关度以及数据流所经过的节点与时间的相关系数,从而准确确定目标数据流所处的节点。通过仿真实验结果和数据分析表明,增量子空间数据挖掘算法的数据挖掘精度在节点、社区数较多的情况下仍达到了较高的挖掘精度。

In order to improve the accuracy of large-scale data-stream mining in complex networks, this paper proposed an incremental subspace data mining method based on data stream density of complex network. It firstly analyzed the data stream density of complex network in the algorithm, and then divided the data stream density of different networks into different communities, so that adopted the undirected traversing loop to determine their corresponding communities of data-flow. It also adopted the incremental subspace data-mining algorithm to calculate the correlation between community network and data-stream and the correlation coefficient between data-stream node and time so as to accurately determine the nodes of the object data stream. Data analysis and simulation experiment results demonstrate that the incremental subspace data mining algorithm still has a higher mining accuracy in the case of a larger number of nodes and communities.