基于网络数据的学术图书评价研究

发布时间：2018-03-16 19:04

本文选题：学术图书评价　切入点：网络数据　出处：《南京大学》2013年硕士论文　论文类型：学位论文

【摘要】：当前学术图书是传承学术思想的重要载体。对于学术界来说,学术图书是学术交流系统的重要组成部分。因此对学术图书的评价工作对于学术研究的参考引用、著者评价和学科知识传承具有相当重要的现实意义。而目前对学术图书的评价方法与期刊评价类似,主要以定性和定量方法相结合进行评价。定性分析主要采用专家评审；定量方法目前还主要是利用与期刊评价、论文评价相同的方法来进行评价,主要依靠一些传统的评价指标,如借阅量、被引量等指标,因而一本图书的质量主要通过其在引文网络中的地位和销售数据等来反映,这显然并没有全面的反映学术图书的质量。另一方面,互联网的发展使得网络中拥有海量的各类数据,以前囿于数据的不可获取性,我们只能借助于少量的数据进行评价工作,现在互联网的快速发展,尤其大数据时代的到来,让我们得以在数据来源和信息技术两方面取得突破性进展,这也意味着我们需要在科研思路上发生重要的思维变革,以更好的抓住时代发展脉络,以更先进、更科学的方法进行学术评价工作。当前,用户对学术图书的评论等很多都发生在网络中,这些数据在学术图书的评价上大有可为。而如何利用这些散落在互联网中的数据并设计相应的评价指标将是本文的研究重点。本文的研究工作主要包括以下几个章节：第一章,绪论。本章阐述了本文的研究背景和意义、研究内容与本文的创新点,以及本文的组织结构。第二章,文献综述。本章首先对现有的学术图书评价指标进行了归纳总结,详细阐述了评价指标的设计原则及具体的指标内容。接着,对现有的学术图书评价研究方法进行归纳和总结,详细介绍了不同方法的应用场景和各自的优势与劣势。第三章,学术图书评价网络数据来源。本章从两个方面对本文的研究进行了数据的准备与描述。首先选取了四个学科在CSSCI中被引次数最高的若干学术图书作为本文的研究样本,其次,详细介绍了目前互联网中主要的图书书评来源网站,对各类型网站进行了详细介绍,并着重阐述了网络数据可用于学术图书评价的原因和意义。第四章,基于Google Scholar引文数据的学术图书评价。Google Scholar引用数据反映了学术图书在学术圈内的影响,本章基于Google Scholar引用数据提出了一组评价指标,包括Goolge Scholar引用次数、Goolge Scholar二级引用次数和Goolge Scholar被引半衰期三个指标,并通过与传统CSSCI被引次数指标的相关性分析证明了新指标的合理性。第五章,基于搜索引擎日志的学术图书评价。搜索引擎是互联网最大的流量入口,用户在搜索引擎中的搜索行为反映了学术图书的网络影响力。本章基于搜索引擎的日志数据提出了一组指标,包括图书被搜索次数、图书搜索被点击次数和图书被页面提及次数三个指标,并通过必应搜索引擎的日志收集指标数据,利用其与传统的图书馆借阅次数指标的相关性分析证明了新指标的合理性。第六章,基于情感挖掘技术的网络评论文本的情感分析评价指标。随着Web2.0的兴起,互联网中散落着许多用户对学术图书的评论,这些评论蕴含着用户的情感倾向,本章在前人研究基础上提出了一种新的基于文本分类的情感挖掘算法。并利用该算法对豆瓣,当当等网站的评论、书评进行情感倾向性判断,这些情感倾向反映了学术图书的网络口碑,进而反映了图书质量。第七章,基于网络数据的学术图书评价体系与系统架构设计。本章结合传统指标和本文提出的基于网络数据的指标提出了一个学术图书评价的指标体系,将指标分为引用型指标,传播型指标,影响力指标和口碑指标四个方面。并设计了一个学术图书的自动化评价系统架构,详细设计了功能模块和技术细节。第八章,结语。本章在前文的基础上总结了本论文的主要结论,并指出了其中存在的不足,为下一步研究工作指明了方向。
[Abstract]:The current academic book is an important carrier of inheriting the academic thoughts. In the academic circles, the academic library is an important part of the scholarly communication system. So the evaluation of academic library work of academic research for reference, has very important practical significance, evaluation and knowledge inheritance. And at present the method of evaluation of academic books and journals the evaluation of similar, mainly by qualitative and quantitative method of combining evaluation. Qualitative analysis mainly adopts expert evaluation; quantitative method is mainly evaluated by the evaluation and periodical, the same method to evaluate, mainly rely on the traditional evaluation indexes, such as loan amount, index citation amount, so the quality of the the book mainly through its citation network and the status of the sales data to reflect the quality of this apparently does not fully reflect the academic books.
On the other hand, the development of the Internet makes all kinds of data in the network has a massive, previously not limited by data availability, to evaluate the work in a small amount of data we can only by now, the rapid development of the Internet, especially in the era of big data, so that we can make a breakthrough in two aspects of data sources and information technology and this also means that we need to have the important change of thinking in scientific research, in order to better grasp the era of development, to more advanced, more scientific methods for evaluation of academic work. At present, users of academic book reviews are occurring in the network, these data have a brilliant future in the evaluation of academic books. And how to use these scattered in the Internet data and design the corresponding evaluation index will be the focus of this article.
The research work of this paper mainly includes the following chapters:
The first chapter, introduction. This chapter describes the background and significance of this study, the content of the research and the innovation of this article, as well as the organizational structure of this article.
The second chapter, literature review. The first chapter of academic books of the existing evaluation index were summarized, elaborated the design principle of evaluation index and specific index. Then, the study of academic books existing evaluation methods and summarized, introduces the application scenarios of different methods and their advantages and disadvantages.
The third chapter, the academic book evaluation network data source. This chapter of this paper from the two aspects of the preparation and description data. First selects four subjects in CSSCI cited several academic books of the highest frequency as the research sample, this paper then introduced the main source of Internet book review for all types of Web sites, is introduced in detail, and focuses on the network data can be used to reason and significance of academic book evaluation.
The fourth chapter, the Scholar.Google Google Scholar Academic Book Evaluation citation data reference data reflects the influence of academic books in academic circles based on this chapter Google Scholar reference data is proposed based on a set of evaluation indicators, including the Goolge Scholar Goolge Scholar two citations, citations and Goolge Scholar cited half-life of three indicators, and through the correlation analysis and the traditional CSSCI index was cited to prove the rationality of the new index.
The fifth chapter, the academic library evaluation of search engine based on log. The search engine is the biggest Internet traffic entrance, user search behavior in the search engine in the network reflects the influence of academic books. This chapter log data search engine is proposed based on a set of indicators, including books by the number of search, book search by clicks and books the page was referred to the number of the three indicators, and through the Bing search engine log collect index data, using the traditional library and the number of indicators related analysis proves the rationality of the new index.
The sixth chapter, the network review text sentiment mining sentiment analysis based on the evaluation index. With the rise of Web2.0, the Internet is littered with many users comment on the academic books, these comments contain user sentiment, this chapter on the basis of previous studies proposed a mining algorithm based on new text classification emotional. Using the algorithm of watercress, Dangdang and other website reviews, book reviews of sentiment analysis, which reflects the emotional tendency of academic books and online word-of-mouth, reflects the quality of books.
The seventh chapter, the academic library network data evaluation system and system based on architecture design. This chapter combines the traditional index and the proposed network data index is proposed based on an index system of academic library evaluation, the index is divided into reference index, spread index, influence four aspects of force index and reputation index. And the design of the automatic evaluation system framework for an academic book, the detailed design of the function module and technical details.
The eighth chapter is the conclusion. This chapter summarizes the main conclusions of this paper on the basis of the foregoing, and points out the shortcomings, pointing out the direction for further research.

【学位授予单位】：南京大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：G353.1

【参考文献】