基于粗糙集属性依赖度强化的交互式大数据特征分类  

Interactive big data feature classification based on attribute dependency enhancement of rough set

作　　者：曹夏琳 CAO Xialin(Xiamen College of Marine Vocational and Technical,College of Information Engineering,Xiamen Fujian 361000)

机构地区：[1]厦门海洋职业技术学院信息工程学院,福建厦门361000

出　　处：《宁夏师范学院学报》2023年第1期90-97,共8页Journal of Ningxia Normal University

基　　金：2019年福建省中青年课题项目(JAT191051).

摘　　要：为了提高交互式大数据特征分类性能,提出基于粗糙集属性依赖度强化的特征分类方法.利用交互式大数据在粗糙集的属性依赖度,确定数据属性之间的关联关系.通过定义近似集强化交互式大数据粗糙集属性依赖度,明确数据状态量之间的幅值规律,并计算出交互式大数据之间的密度值.根据密度值确定交互式大数据分布的集中区域并提取其特征,构建目标函数计算出大数据的期望熵值.利用交互式大数据特征的增益信息值,实现特征分类.实验结果表明,文中方法在分类交互式大数据特征时,AUC值接近1.0,具有较高的应用价值;文中方法分类时的加速比高于4.0,提高了分类效率.In order to improve the performance of interactive big data feature classification,a feature classification method based on rough set attribute dependency enhancement is proposed.Use the attribute dependency of interactive big data in rough set to determine the association relationship between data attributes.By defining the approximation set to strengthen the attribute dependency of interactive big data rough set,define the amplitude rule between data state variables,and calculate the density value between interactive big data.According to the density value,determine the concentrated area of interactive big data distribution and extract its characteristics,and construct an objective function to calculate the expected entropy value of big data.Use the gain information value of interactive big data features to realize feature classification.The experimental results show that the AUC value of the proposed method is close to 1.0 when classifying interactive big data features,which has high application value.The acceleration ratio of the method in this paper is higher than 4.0,which improves the classification efficiency.

关 键 词：粗糙集 特征提取 特征分类 依赖度强化 交互式大数据 属性值 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]