文章摘要

李广建,江信昱.论计算型情报分析[J].中国图书馆学报,2018,44(2):4~16
论计算型情报分析
On Computational Information Analysis
投稿时间:2017-10-11  
DOI:
中文关键词: 情报分析  大数据  计算型情报分析  计算化  深度分析
英文关键词: Information analysis  Big data  Computational Information Analysis  Computability  Advanced analytics
基金项目:本文系国家社会科学基金重点项目“大数据环境下的计算型情报分析方法与技术研究”(编号:14ATQ005)的研究成果之一
作者单位E-mail
李广建 北京大学信息管理系 北京 100871 ligj@pku.edu.cn 
江信昱 北京大学信息管理系 北京 100871  
摘要点击次数: 696
全文下载次数: 
中文摘要:
      受大数据环境和第四范式的影响,情报分析走向计算的趋势已经越来越明显。在信息设备低廉化、处理对象数字化、计算方法智能化的共同作用下,情报分析逐步走向计算型情报分析。计算型情报分析是情报分析与以计算机技术为核心的信息技术相结合的产物,涵盖目标、功能、实现、分析方式和知识体系五个层面的内容;以计算机及相关信息技术为工具,以机器学习与知识理解等为核心技术,以情报分析方法和数学模型方法为组织分析手段,通过对数据内容及其关系、模式的深度解析、挖掘和发现,帮助分析人员解决情报问题,完成情报任务。计算型情报分析的思维理念包括定量思维、自动化思维、融合思维、容错思维等,在理论、方法和技术、系统、应用实践等四个方面都有诸多研究课题等待我们探索和挖掘。参考文献51。
英文摘要:

    Due to the impact of the big data environment and the fourth paradigm of scientific research, the trend of computability in information analysis has become more obvious. First, as information equipment becomes cheaper, the cost of tools for information analysis decreases and the path for the computability of information analysis widens. Second, most of the objects to be processed are in digital forms which enable a wider span of data resource to be analyzed. This creates more opportunities for computing in information analysis. Third, computing methods have become more intelligent, which brings about the ability to carry out an in depth analysis and provides a deep understanding of the information. These three factors play a decisive role in making the transition from information analysis to Computational Information Analysis(CIA).
CIA is both a theory and an approach which derives from information analysis and information technology of which computer technology is at the center. It is proposed to address the limitation of human's ability in collecting, processing and analyzing large amount of information. In CIA, computers and related information technology are adopted as basic tools and intelligent technologies like machine learning and knowledge understanding are incorporated as core technology. Information analysis methods and mathematical models are used to organize and analyze information. Through a deeper analysis and mining in the content, relationship and pattern in the data, CIA helps analysts solve analysis problem and tasks and make decisions.
While human is inadequate in gathering, processing and analyzing large amount of data, CIA makes up for that shortcoming with the ability to harness the fast computing power of computers and is also more objective when compared with human. CIA provides a more comprehensive, faster, accurate, objective and deeper analysis result for the target problems thanks to its advantages in harnessing the power of computer technology and quantitative analysis methods. To be more specific, CIA is driven by data mining and machine learning techniques on the analysis environment which is built upon computers. The analysis of CIA is mainly quantitative and places an emphasis on methods that can be processed by machines to undergo quantitative, computational and testable process. Besides, CIA also contributes to the knowledge system as it not only inherits theories and methods from information analysis but also borrows theories and methods from computer science, data science, mathematical statistics and other fields and develops new theories and methods with computational characteristics.
CIA has its own thinking and ideas. Quantitative thinking is to abstract quantitative mathematical forms from objective things and phenomena. Automatic thinking is to transform data and analysis method in a form that can be processed by a computer. Integrative thinking places an importance on the relevance of different data sources. By fusing multiple data sources, it is possible to not only reveal correlations but also complement or cross validate using different kinds of data. Fault tolerance thinking means that precision loss at a micro level and a certain amount of errors and confusions is tolerable.
Finally, this paper raises research topics which deserve more attention under different perspectives, i.e. theory, method and technique, system, and application practice, including but not limited to: constructing a theory of data advanced analytics based on the integration of information science theory and related disciplines, discussing on automatic implementation and improvement of different information analysis methods in CIA, designing an intelligence analysis system within task data method which is interrelated, combining with new computing technology or multiple data sources for deep data content mining, etc. 51 refs.

下载全文   查看/发表评论  下载PDF阅读器