一种基于Spark的大规模语义数据分布式推理框架
发布时间:2018-02-24 17:04
本文关键词: Spark 并行化语义推理 分布式框架 语义大数据 出处:《计算机科学》2016年S2期 论文类型:期刊论文
【摘要】:随着大规模语义数据的涌现,研究高效的并行化语义推理成为热点问题之一。现有推理框架大多存在可扩展性方面的不足,难以满足大规模语义数据的需求。针对现有推理框架的不足,提出一种基于Spark的大规模语义数据分布式推理框架。该框架主要包括语义建模、规则提取和基于Spark的并行推理机等3个模块。通过过程分析和推理实例验证,提出的分布式并行推理的计算性能(T(n)=O(log_(2)n))远远优于顺序式推理的计算性能(T(n)=O(n))。
[Abstract]:With the emergence of large-scale semantic data, the research of efficient parallel semantic reasoning has become one of the hot issues. It is difficult to meet the needs of large-scale semantic data. A distributed reasoning framework for large-scale semantic data based on Spark is proposed, which mainly includes semantic modeling. Rule extraction and parallel inference engine based on Spark are three modules. Through process analysis and reasoning case verification, the computational performance of distributed parallel reasoning is much better than that of sequential reasoning.
【作者单位】: 大连外国语大学软件学院;
【基金】:国家自然科学基金项目(61371090) 辽宁省自然科学基金项目(2015020017) 大连外国语大学校级科研项目(2014XJQN09)资助
【分类号】:TP311.13
,
本文编号:1531018
本文链接:https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/1531018.html