图划分的谱聚类算法在命令票系统校核文本挖掘中的应用

The Application of Spectral Clustering Algorithm of Graph Partition in Checking Text Mining of Command Ticket System

由于电力调度系统中命令票内容的不同,现有文本挖掘方法难以得到精准结果,因此提出基于图划分谱聚类算法的命令票系统校核文本挖掘方法。根据命令票内容将其分为不同的种类,引入二部图建模分析各个组间的关系;利用谱聚类算法中Laplancian矩阵做图分割处理,获得矩阵的特征向量和数据点集合,凭借维数特征数据集完成得到聚类结果,即命令票校核文本挖掘结果。实验结果表明,所提方法的挖掘准确性在80%以上,挖掘时间最高19.5 ms,稳定性保持在80%以上,符合电力企业需求。

Due to the different contents of command tickets in the power dispatching system,a method is proposed to improve the accuracy of ticket contents.The method is divided into different types according to the contents of the command ticket,in-troduces bipartite graph model to analyze the relationship between each group,uses the Laplacian matrix in the spectral cluste-ring algorithm to obtain the feature vectors and data points of the matrix,and complete the dimension feature dataset to check the text mining results.The experimental results show that the mining accuracy of the proposed method is above 80%,the highest time is only 19.5 ms,and stability of the mining time remains above 80%,which meets the needs of power enterprises.