基于分布式数据的学习分类器的研究  

Research on Learning Classifier Based on Distributed Data

作　　者：信晓艺 XIN Xiao-yi(School of Mathematics and Big Data,Dezhou University,Dezhou,253023,Shandong)

机构地区：[1]德州学院数学与大数据学院,山东德州253023

出　　处：《蚌埠学院学报》2022年第2期76-80,共5页Journal of Bengbu University

摘　　要：大数据的应用催生出分布式数据,分布式数据中对象或个体表示特征值单位组的K元组,而每个单位组的特征值取样又源于特定特征和特定对象的分布式数据。从分布式数据角度来研究和解决学习分类器的问题,基于分布式数据生成模型的方法和基于分布式数据生成模型的辨别方法进行分析,比较了三种算法对真实及合成分布式数据集的表现。实验结果表明,利用以分布式事例形式表现的信息的分类器明显优于那些未充分利用此类信息的方法。The application of big data has given birth to distributed data.In distributed data,objects or individuals represent K-tuples of feature value unit groups,and the feature value sampling of each unit group is derived from the distributed data of specific features and specific objects.It proposed to study and solve the problem of learning classifiers from the perspective of distributed data in this study,the method based on the distributed data generation model and the discrimination method based on the distributed data generation model were analyzed,and the performance of the three algorithms on real and synthetic distributed data sets was compared.The results showed that classifiers that use information in the form of distributed cases are better than those that underutilize such information.

关 键 词：分布式数据 学习分类器 分布式事件 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]