摘要
为了提高海量档案高维特征的超大数据集管理及分析效率,在研究了NGN、5G网络、多核SVM等相关理论及概念基础上,设计了一种智能档案管理系统。该架构按层级可分为应用层、控制层和基础设施层,系统协调作用实现了高效地档案数据管理及分析功能。其中,提出了一种混合数据分析方法以提高数据分析效率。从大规模的未标记数据集中选择具有代表性的实例,以减少标记工作和训练时间,实现了不损失精度情况下提高数据分析效率。通过仿真及分析,与传统SVM方法相比,在不明显降低准确率情况下,提出的方法只需要标记一小部分代表性的训练实例可完成训练过程。
In order to improve the efficiency of large data set management and analyze of high-dimensional features of massive archives, based on the research of NGN, 5G network, multi-core SVM and other related theories and concepts, this paper designs an intelligent archives management system. The architecture can be divided into application layer, control layer and infrastructure layer. The coordination function realizes the efficient file data management and analysis function. In order to improve the efficiency of data analysis, a hybrid data analysis method is proposed, it selects representative examples from large-scale unlabeled data sets to reduce the marking work and training time, and improves the efficiency of data analysis without losing accuracy. Through simulation and analysis, compared with the traditional SVM method, this method only needs to mark a small number of representative training examples to complete the training process without significantly reducing the accuracy.
作者
鲜娅静
XIAN Yajing(Discipline Inspection Commission,Xi’an Medical University,Xi’an 710021,China)
出处
《微型电脑应用》
2023年第2期161-163,168,共4页
Microcomputer Applications
