面向数字人文研究的大规模古籍文本可视化分析与挖掘

传统的古籍开发与应用模式已难以适应人文学科研究的需要,人文学科研究者期待一个技术逻辑和人文逻辑相耦合的数字人文研究范式的出现。本文从古籍文献深层次开发与利用出发,利用新的信息技术与面向数字人文研究跨学科方法,以大规模中国古籍文本为研究对象,采用大数据研究理念,对古籍进行整理、标注、自动分词等处理,以词频分析统计为研究核心,采用数据降噪、基于窗口时间单位的统计分析计算、滑动窗口预测等分析与挖掘方法,采用大数据实时分析技术,实现了实时、在线、立体、可视化、定量分析字词的历史词频分布规律,创建了一个以语言学、历史文献学、历史地理学等人文学科研究为主的古籍实时统计分析平台,可辅助研究者在大量的古籍文献中发现新的模式、现象、趋势等,实现古籍开发与应用模式创新的初步尝试。图11。参考文献36。     

Web数据关联创建策略研究

调研关联数据的关联关系创建算法和策略,分析同构模式下属性相似度和图形相似度算法以及相应的组合策略,对比研究异构模式下包含和不包含实例信息的架构映射方法,对可创建丰富语义关联的推导传递的思想进行剖析,并提出关联创建面临的挑战。     

“可获得性论”的新启航

总结了“可获得性论”独特之处以及围绕这个理论的学术争鸣,指出Web2．0时代对图书馆学基础理论研究的挑战。从可获得性论的基本过程出发,认为可获得性论原定义假设用户需要的知识单元不言自明,是不确切的。提出“用户需要的知识单元在文献群中的可获得性”,作为可获得性论的新定义,并对新定义作了一些基本分析。参考文献4。     

档案网站调查报告的现状分析与建议

调查研究是制定正确的政策、方案、标准的前提。调查过程可以认识调查对象的现状,调查结果可以发现问题的根本原因所在,从而实现调查的目的:提出建设性意见或制定出有效措施,使问题得以解决,推动事物的向前发展。本文拟对档案网站调查报告进行分析与研究,了解档案网站调查报告的现状和问题,对以后调查报告的框架结构、调查的形式和方法、内容的设置,提出合理化建议,为网站调查报告的模式构建提供实践数据,为帮助调查报告水平的提高、为建立国家档案网站统一的评价标准提供物质基础,为档案网站相关方面的研究提供参考。     

Articles vs. proceedings papers: Do they differ in research relevance and impact? A case study in the Library and Information Science field

In 2008, the type of document “proceedings paper” (PP) was assigned in the WoS database to journal articles which were initially presented at a conference and later adapted for publication in a journal. Since the use of two different labels (“article” and “proceedings paper”) might lead to infer differences in their relevance and/or quality, this paper presents a comparative study of standard journal articles and PP in journals to explore potential differences between them. The study focuses on the Library and Information Science field in the Web of Science database and covers the 1990–2008 period. PP approximately account for 9% of the total number of articles in this field, two-thirds of which are published in monographic issues devoted to conferences, which tend to be concentrated in specific journals. Proceedings papers emerge as an heterogeneous set comprising PP in ordinary issues, similar to standard articles in structure and impact of research; and PP in monographic issues, which seem to be less comprehensive and tend to receive less citations. Faster publication of PP in monographic than in ordinary issues may conceal differences in the review process undergone by either type of paper. The main implications of these results for authors, bibliometricians, journal editors and research evaluators are pointed out.     

IP Address Data and its Place in the Library Community Today and in the Future

Internet Protocol (IP) technology has played an important role in the management of access to scholarly work within institutions since the advent of electronic publishing in the late 1990s. Despite recent efforts to explore alternative modes of authentication, the ubiquitous use of IP address authentication within the systems of libraries, publishers, and third party vendors around the world suggests it will continue to play a significant role for the foreseeable future. However, recent studies show that a large amount of the IP address data held within publisher systems is incorrect, improperly assigned, or requires clarification. Poor source data carries significant risks—not the least of which translates into unreliable usage reporting for libraries and publishers. The IP Registry is designed to help mitigate these challenges by creating a Registry of cleaned data that libraries and publishers can interact with in a centralized fashion thereby helping to ensure accuracy and reducing the level of resources necessary to manage and communicate changes to existing data. The database of more than 60,000 academic, corporate, and government libraries was created through the amalgamation of raw source data from more than 170 scholarly publishers. In addition to cleaned IP address data, the Registry includes standardized English naming, aliases, local language naming (and characters), hierarchies, and organizational IDs for each record. The importance of accurately understanding current usage and how libraries utilize published content is clear. Regardless of what the future holds, the need for a clean and stable source of IP address data is real for libraries, publishers, and vendors alike.     

面向Web3.0的分众分类研究

在全面分析Web3.0及分众分类的基本特征的基础上,针对Web3.0的的核心理念和目前分众分类存在的局限性,从微内容的自由整合与过滤、微内容的标准化与结构化、标注的规范控制、标签的基本属性、标签分类与聚类、标签间的关联、标签与本体的结合等方面探讨面向Web3.0的分众分类未来研究方向,最终实现分众分类体系的优化。     

Bibliometric profile of top-cited single-author articles in the Science Citation Index Expanded

In this study, we identified and analyzed characteristics of top-cited single-author articles published in the Science Citation Index Expanded from 1991 to 2010. A top-cited single-author article was defined as an article that had been cited at least 1000 times from the time of its publication to 2012. Results showed that 1760 top-cited single-author articles were published in 539 journals listed in 130 Web of Science categories between 1901 and 2010. The top productive journal was Science and the most productive category was multidisciplinary physics. Most of the articles were not published in high-impact journals. Harvard University led all other institutions in publishing top-cited single-author articles. Nobel Prize winners contributed 7.0% of articles. In total, 72 Nobel Prize winners published 124 single-author articles. Single-authored papers published in different periods exhibited different patterns of citation trends. However, top-cited articles consistently showed repetitive peaks regardless of the time period of publication. “Theory (or theories)” was the most frequently appeared title word of all time. Leading title words varied at different time periods, and only five title words, method(s), protein(s), structure(s), molecular, and quantum consistently remained in the top 20 in different time periods.     

基于Fiddler代理程序的电子资源使用统计分析系统的设计与应用

[目的/意义]对比分析国内图书馆自行采集电子资源使用统计数据的主要方法,解决现有技术方法无法采集基于HTTPS协议的电子资源访问信息的问题。[方法/过程]在现有基于旁路监听的采集模式的基础上对技术加以改进,通过策略路由和Fiddler代理程序来实现基于HTTPS协议访问的电子资源使用数据的统计和分析,并在此基础上设计并应用电子资源使用统计分析系统。[结果/结论]解决了基于HTTPS协议访问的电子资源使用数据的采集问题,对其他图书馆自行采集电子资源使用统计数据有一定的借鉴作用。     

国内外专利挖掘研究(2005-2014)综述

在中国知网、万方数据和Web of Science进行检索,获得72篇相关中文文献和98篇英文文献,并从中选出66篇进行综述。专利挖掘研究包括术语抽取、聚类、分类、以复杂网络为基础的方法、以时间为基础的方法和基于专利挖掘的技术研究等6个主题。尽管近10年来这一领域发展较快,但是部分研究也存在试验验证不精确、基于IPC的自动分类效果不好、所要解决的问题不明确且局限于方法应用和粒度粗糙等问题。专利挖掘研究应该注重发现问题,而非简单地应用方法。