CAJ | 학술논문

为有效精确地挖掘海量数据流特征，提出采用数据集中位分割和冗余数据碎片合并的方法设计决策树并构建云平台数据特征挖掘模型。传统的云平台下数据挖掘中对碎片信息不做处理，使文本碎片成几何级增长，导致有用信息的丢失。通过挖掘冗余信息中符合挖掘条件的碎片特征，使用KD树进行数据挖掘索引，在迭代过程中，对数据进行层进中位分割，并结合碎片合并技术，构建数据挖掘云平台模型，对中位数进行维度匹配分箱，使数据的挖掘和传输率最大限度地得到利用。仿真实验证明了采用碎片合并的方法能有效提取冗余碎片中的有用特征信息，数据挖掘性能得到大幅度提升，在数据信息提取和管理中具有很好的推广意义。
위유효정학지알굴해량수거류특정，제출채용수거집중위분할화용여수거쇄편합병적방법설계결책수병구건운평태수거특정알굴모형。전통적운평태하수거알굴중대쇄편신식불주처리，사문본쇄편성궤하급증장，도치유용신식적주실。통과알굴용여신식중부합알굴조건적쇄편특정，사용KD수진행수거알굴색인，재질대과정중，대수거진행층진중위분할，병결합쇄편합병기술，구건수거알굴운평태모형，대중위수진행유도필배분상，사수거적알굴화전수솔최대한도지득도이용。방진실험증명료채용쇄편합병적방법능유효제취용여쇄편중적유용특정신식，수거알굴성능득도대폭도제승，재수거신식제취화관리중구유흔호적추엄의의。
In order to mining massive data flow features efficiently and accurately, middle data segmentation and redundant data fragments were proposed to design decision tree and build a cloud platform data mining model. KD tree was used for data mining index. In the iterative process, the data was divided into a layer, and the median split box matching dimension was processed. It made the mining and data transmission rate improve greatly. Simulation result shows that the useful frag-ments and feature information can be mining effectively, the data mining performance can be improved greatly, and it has good popularization significance in information extraction and data.management.