基于链接聚类的OA期刊网络影响因子研究

发布时间：2018-03-03 22:13

本文选题：OA期刊　切入点：网络影响因子　出处：《南京农业大学》2010年硕士论文　论文类型：学位论文

【摘要】：开放获取期刊(也称OA期刊)自发行以来,倡导了学术共享的理念与出版机制,以免费阅读、下载、复制、散发或链接等优势越来越受到广大专家学者的青睐。在大力发展和使用OA期刊的同时,人们也逐渐关注OA期刊的评价问题。OA期刊有其自身的链接特点,且不同链接类型的价值也不同,因此以往的网络影响因子研究方法已不能满足人们科学评价OA期刊的需求。OA期刊网络影响因子的测算需要一种新的结合OA期刊链接特点的解决方案。本文结合OA期刊的链接特征,提出了一种基于链接聚类的OA期刊网络影响因子的计算方案。为了划分链接类型,寻找各个链接之间隐含的相同的链接动机,作者利用内容特征、出链特征、入链特征为链接到某OA期刊的每个网页构建了三元组向量,并使用M-S算法进行聚类分析。在聚类算法选取的基础上,作者利用Java语言设计了专门的链接抓取工具和聚类分析工具用以实现聚类分析过程,综合使用数据库技术与专门编写的程序进行数据清洗。在链接聚类分析的基础上,文章从链接分析的假设前提入手,将链接类型与链接权重因素加入到网络影响因子的计算中,提出了基于链接聚类的OA期刊网络影响因子计算公式。文章根据基于链接聚类的OA期刊网络影响因子计算方法,利用专门设计链接抓取工具和聚类计算工具,对D-Lib Magazine的站外链接进行了聚类分析,计算出了D-Lib Magazine基于链接聚类的网络影响因子值,并与传统网络影响因子计算方法进行了比较。实证结果表明,基于链接聚类的OA期刊网络影响因子的计算方法是可行的,能够结合OA期刊链接特点评价OA期刊。文章提出的基于链接聚类的OA期刊网络影响因子研究方法,可以为OA期刊评价服务,引导大众阅读,促进国内OA期刊提升办刊质量。
[Abstract]:Since its publication, Open access Journals (also known as OA Journals) have advocated the idea of academic sharing and publishing mechanisms for free reading, downloading and copying. The advantages of distributing or linking are more and more popular among experts and scholars. With the development and use of OA periodicals, people pay more and more attention to the evaluation of OA periodicals. OA periodicals have their own link characteristics. And different types of links have different values, Therefore, the previous research methods of network impact factors can not meet the needs of scientific evaluation of OA journals. The calculation of network impact factors of OA periodicals requires a new solution combining the characteristics of OA periodicals. Based on the link characteristics of OA journals, this paper proposes a scheme to calculate the network influence factors of OA periodicals based on link clustering. The author uses content feature, chain feature, chain feature to construct triple vector for each web page linked to a certain OA periodical, and uses M-S algorithm to cluster analysis. In order to realize the cluster analysis process, the author designed a special link grab tool and cluster analysis tool by using Java language, and used the database technology and the specially written program to clean the data. On the basis of the link clustering analysis, the author used the database technology and the specially written program to clean the data. This paper begins with the hypothetical premise of link analysis, adds the link type and link weight factors to the calculation of network influence factors, and puts forward a formula for calculating the network influence factors of OA periodicals based on link clustering. According to the method of calculating the influence factors of OA periodical network based on link clustering, the paper makes cluster analysis of the out-of-station link of D-Lib Magazine by using the specially designed link grab tool and clustering calculation tool. The network influence factor value of D-Lib Magazine based on link clustering is calculated and compared with the traditional network influence factor calculation method. The empirical results show that the calculation method of network influence factor based on link clustering is feasible. Can combine OA periodical link characteristic to evaluate OA periodical. The research method of network influence factors of OA periodicals based on link clustering can serve the evaluation of OA periodicals, guide the public to read, and promote the quality of OA periodicals in China.
【学位授予单位】：南京农业大学
【学位级别】：硕士
【学位授予年份】：2010
【分类号】：G237.6;TP311.13

【参考文献】