摘要
制定重大决策需要全方面的数据支持,而大数据时代的到来却让决策信息搜集更加困难。数据多样性和价值密度低的特点,让决策者耗费大量时间在信息需求的表达上。设计了一种通用的信息需求描述语言,及其为数据库、发布/订阅、搜索引擎等常见信息资源解析转换的方法,使得用户可以通过一种统一的方式将相关异构信息全部搜集到。实验表明,其基于本体的语义描述能力能够有效屏蔽数据异构性,提高信息搜集的全面性和准确性。
Important decision making requires all-source data support. However, in the age of big data, collecting decision information becomes even more difficult due to data variety and heterogeneity. To get precise search results, users spend much time on information requirement expression, which causes low efficiency. We design a general language for information requirement description, along with its translation methods for common resources such as databases, publish-subscribe systems and search en- gines. Thus all-source heterogeneous information can be collected by one request in a unified format. Test results demonstrate that its ability to express requirement semantics can solve the cross-area heterogeneity problem of big data with the aid of ontology techniques, and therefore improves the precision and comprehension of information collection.
出处
《计算机工程与科学》
CSCD
北大核心
2015年第8期1436-1443,共8页
Computer Engineering & Science
基金
国防973计划资助项目
关键词
大数据
决策支持
信息搜集
需求描述
big data
decision support
information collection ~ requirement description
