基于混合频繁模式树的粗糙集属性约减算法的研究与应用  

Research and application on rough set attribute reduction model based on mixed frequent pattern tree

作　　者：林春喜 徐宏喆[1] 王谊青 李文[1] Lin Chunxi;Xu Hongzhe;Wang Yiqing;Li Wen(School of Electronics&Information Engineering,Xi’an Jiaotong University,Xi’an 710049,China;School of Data Science&Computer Science,Sun Yat-sen University,Guangzhou 510006,China)

机构地区：[1]西安交通大学电子信息工程学院,西安710049 [2]中山大学数据科学与计算机学院,广州510006

出　　处：《计算机应用研究》2018年第4期988-991,1027,共5页Application Research of Computers

摘　　要：粗糙集对于学习分析系统的属性约减模型有着重要的研究意义和使用价值。针对教育大数据高维度、不完备、增量性等现状,提出了基于不完备决策表的差别信息增量更新算法,并结合树型结构对差别信息的高效存储和粗糙集的核属性概念,设计构建了MIX_FP树,实现高维属性的有效约减。实验结果验证了该算法具有较好的运行效率和空间性能,为教育大数据的属性约减提供了有效的方法,同时为基于粗糙集理论的属性约减算法研究及其在学习分析领域的应用提供了新的研究思路。Rough set theory has essential meaning in theory and practical value for the attribute reduction model of learning analysis system.In order to solve the problem of the high dimension,incompleteness and incremental property of education big data,this paper proposed a discernibility information algorithm and an incremental updating algorithm based on incomplete decision table.Then,it combined the tree structure for effective storage of discernibility information and the core concept of rough set to design and construct the MIX_FP tree(mixed frequent pattern tree)structure,which could provide effective reduction of high dimension attribute of information system decision table.The experiment results show that the rough set attri-bute reduction algorithm based on MIX_FP tree has good operational efficiency and spatial performance.The model provides an effective support for the attribute reduction of education data,and provides a new research idea for the research and application on learning analysis field of attribute reduction algorithm based on rough set theory.

关 键 词：属性约减 粗糙集 差别信息 MIX_FP树 学习分析技术 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]