粗糙集下网络大数据混合属性特征检测仿真

由于网络混合属性集的冗余数据量多,影响数据检测的查全率,为此提出结合粗糙集理论的网络大数据混合属性特征检测方法。首先构建一个四元组,利用四元组的任意邻域信息测算其长度函数,以判断信息特征的相似性,结合粗糙集理论求解相似信息特征的邻域熵,以检测并分类重复数据属性。为优化数据分类效率,引入支持向量机分类思想,将大数据混合属性的分类问题变换为线性可分问题,实现网络大数据混合属性特征检测与分类。实验结果表明,所提方法能够有效根据数据特征筛选出无关数据信息,使用经过训练后的分类装置对约简后的特征集进行分类,与基于特征和分类器参数组合优化的网络属性特征检测方法比较,证明了所提方法的有效性,为网络大数据混合特征检测技术提供一种新的有效解决方式。

Simulation of Mixed Attribute Feature Detection for Large Network Data Based on Rough Set

Due to large amount of redundant data in the mixed attribute set,the recall rate of data was low.Therefore,a method to detect mixed attribute feature of network big data based on theory of rough set was put forward.Firstly,a quaternion was constructed,and its length function was measured by any neighborhood information of quaternion,so that the similarity of information features was determined.Secondly,the neighborhood entropy of similar information features was solved by combining with the rough set theory,so that the duplicate data attributes was detected and classified.In order to optimize the data classification efficiency,the classification idea of support vector machine was introduced,and the classification problem of big data mixed attributes was transformed into a linear divisible problem.Thus,the feature detection and classification of big data mixed attributes in network were realized.Simulation results verify that the proposed method can effectively find out the independent data by data features,and use the trained classification device to classify the reduced feature sets.Compared with the method of network attribute feature detection based on the parameter combination optimization of features and classifiers,the effectiveness of the proposed method can be proved,so the proposed method can provide a new and effective way for the network big data mixed feature detection technology.