基于梯度提升回归树的大数据集离群点挖掘模型构建  

Construction of Big Data Cluster Point Mining Model Based on Gradient Lifting Regression Tree

作　　者：胡小琴 HU Xiao-qin(Quanzhou University of Information Engineering, Quanzhou Fujian 362000,China)

机构地区：[1]泉州信息工程学院

出　　处：《佳木斯大学学报（自然科学版）》2019年第5期743-747,共5页Journal of Jiamusi University：Natural Science Edition

摘　　要：为了提高大数据集离群点挖掘能力,提出基于梯度提升回归树的大数据集离群点挖掘模型,构建大数据集离群点的回归树分布模型,采用多维特征融合方法进行大数据集离群点的特征检测,提取大数据集离群点的空间区域分布特征量,采用梯度提升回归分析方法对提取的大数据集离群点特征进行模糊聚类处理,在聚类中心中实现对大数据集离群点数据的自适应融合和分布式检测,通过梯度提升回归树分析方法实现大数据集离群点挖掘。仿真结果表明,采用该方法进行大数据集离群点挖掘的准确性较高,抗干扰性较好,提高了大数据集离群点挖掘过程的收敛和控制能力。In order to improve the mining ability of big data cluster points, a big data cluster point mining model based on gradient lifting regression tree is proposed, and the regression tree distribution model of big data cluster points is constructed. The multi-dimensional feature fusion method is used to detect the features of big data cluster points, and the spatial regional distribution features of big data cluster points are extracted. The gradient lifting regression analysis method is used to fuzzy clustering the extracted big data cluster point features, and the adaptive fusion and distributed detection of big data cluster point data are realized in the clustering center. Big data cluster point mining is realized by gradient lifting regression tree analysis method. The simulation results show that the method has high accuracy and good anti-interference ability, and improves the convergence and control ability of big data cluster mining process.

关 键 词：梯度提升回归树 大数据 集离群点 挖掘 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]