在线图处理:面向信息网络的在线分析处理
首发时间:2012-05-22
摘要:信息网络是以图为核心的高级数据组织形式,其与传统数据形态的核心区别在于每条记录都包含一个或多个图(或网络)。对信息网络数据集进行面向主题的、多维的、多层次的在线分析处理(OLAP)面临数据规模爆炸、问题复杂性剧增等技术难题,传统OLAP技术不再适用。本文提出面向信息网络的在线分析处理模型OLGP(On-Line Graphic Processing),设计OLGP模型的核心操作(信息维上卷、拓扑维上卷)算法,实现在线图处理器OLGPer1.0。实验表明本文算法能够有效完成合作者网络的多维分析,对含有17431位作者的12499篇ACM论文的真实合作者网络数据集进行相关分析处理,信息维上卷仅需要113毫秒,拓扑维上卷仅需要61毫秒。
关键词: 信息网络 信息网络数据立方 信息网络基方体 在线图处理 异步上卷
For information in English, please click here
On-Line Graphic Processing: Information network oriented On-Line Analytical Processing
Abstract:Information network is an advanced organizational form of data base on graph. The core difference to traditional relational database is each record contains one or more graphic dataset (or network). There remain some technical problems such as data size explosion, dramatic increase in the complexity of problems when dealing with subject-oriented, multi-dimensional and multi-layer On-Line Analytical Processing (OLAP) of the information network dataset. Thus traditional OLAP technology and tools are no longer applicable. The OLGP (On Line Graphic Processing) model based on data set of information network is raised in the paper; then, as the core operating algorithm of the model OLGP, informational roll-up and topological roll-up are designed; at last the On-Line Graphic Processer OLGPer1.0 is implemented. Experiments show that the proposed algorithms can perform multi-dimensional analysis of collaboration network effectively and efficiently. It costs only 113 ms on informational roll-up and only 61 ms on topological roll-up to analyze and process a real collaboration network dataset of 17431 authors and 12499 papers.
Keywords: Information Network (InfoNetwork) InfoNetCube InfoNetBaseCube On-Line Graphic Processing Asynchronous roll-up
基金:
论文图表:
引用
No.****
同行评议
共计0人参与
勘误表
在线图处理:面向信息网络的在线分析处理
评论
全部评论0/1000