基于关联分析的异常信息属性特征提取仿真  

Feature Extraction Feature Extraction of Anomaly Information Based on Association Analysis

作　　者：王哲[1] 徐燕文[1] WANG Zhe;XU Yan-wen(School of Information Technology,Henan University of Chinese Medicine,Henan Zhengzhou 450046,China)

机构地区：[1]河南中医药大学信息技术学院

出　　处：《计算机仿真》2019年第12期406-410,共5页Computer Simulation

基　　金：河南省社科联调研课题(SKL-2019-902)

摘　　要：针对现阶段异常信息属性特征提取方法无法及时对数据库更新的数据进行处理的缺陷,研究并提出一种基于关联分析的异常信息属性特征提取方法。组建无向的频繁模式网络模型,分析模型特点,获得项目关联矩阵;通过自适应盲源分离方法,对信息的相关特征进行融合以及挖掘,得到不同特征值之间的关系,通过特征的分布情况,使用关联规则信息融合方法对项目关联矩阵进行聚类处理,在整个数据的聚类中心,通过计算得到判决阈值,将其与自适应回归方法对异常信息属性特征进行特征挖掘,通过信息融合方法实现异常信息属性特征提取。仿真结果表明,与传统提取方法相比所提方法能够快速、准确提取异常信息属性特征。In this paper,a method to extract abnormal information attribute features based on association analysis was proposed.At first,the network model of undirected frequent pattern was built and the characteristics of model were analyzed to get the association matrix.Through the adaptive blind source separation method,the related information features were fused and mined,so that the relationship between different feature values was obtained.Based on the distribution of features,the association rule information fusion method was used to cluster the association matrix.In the clustering center of data,the decision threshold was calculated.The feature of abnormal information attribute was mined by the decision threshold and adaptive regression method.Finally,the feature extraction of abnormal information attribute was realized by information fusion method.Simulation results show that,compared with the traditional extraction method,the proposed method can extract the attribute features of abnormal information quickly and accurately.

关 键 词：关联分析 异常信息 属性特征提取 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]