社会网络分析教程(研读分享案例-用UCINET软件进行社会网络分析)
社会网络分析教程(研读分享案例-用UCINET软件进行社会网络分析)比如,要研究国内学者在供应链领域的研究热点问题,学者可以选择中国知网作为文献数据库,以“供应链”为关键词或者篇名进行检索,并通过对文献的多次筛选,得到最有参考价值、最具学术水平的文献。通常,在进行社会网络分析之前,最重要的一步是找到原始数据。#2UCINET软件介绍作为最受欢迎的社会网络分析软件之一,UCINET软件常用来进行一维与二维数据分析,利用UCINET软件可以读取文本文件、KrackPlot、pajek、Negopy、VNA等格式的文件。它能处理32767个网络节点。当然,从实际操作来看,当节点数在5000~10000之间时,一些程序的运行就会很慢[2]。#3一个小案例
分享兴趣、传播快乐、增长见闻、留下美好,大家好,这里是LearningYard学苑,今天小编为大家带来文章:案例-用UCINET软件进行社会网络分析.
#1社会网络分析法
社会网络分析法是一种社会学研究方法,社会学理论认为社会不是由个人而是由网络构成的,网络中包含结点及结点之间的关系,社会网络分析法通过对于网络中关系的分析探讨网络的结构及属性特征,包括网络中的个体属性及网络整体属性,网络个体属性分析包括:点度中心度,接近中心度等;网络的整体属性分析包括小世界效应,小团体研究,凝聚子群等[1]。
#2UCINET软件介绍
作为最受欢迎的社会网络分析软件之一,UCINET软件常用来进行一维与二维数据分析,利用UCINET软件可以读取文本文件、KrackPlot、pajek、Negopy、VNA等格式的文件。它能处理32767个网络节点。当然,从实际操作来看,当节点数在5000~10000之间时,一些程序的运行就会很慢[2]。
#3一个小案例
通常,在进行社会网络分析之前,最重要的一步是找到原始数据。
比如,要研究国内学者在供应链领域的研究热点问题,学者可以选择中国知网作为文献数据库,以“供应链”为关键词或者篇名进行检索,并通过对文献的多次筛选,得到最有参考价值、最具学术水平的文献。
通过建立文献作者与文献关键词的矩阵,也就是作者与研究问题的2-mode网络。这就是我们做社会网络分析所需要的原始数据。
下面,我们进行社会网络分析的具体操作。案例中使用的数据来源于参考文献[2]。
第一步:导入原始数据方法一
依次点击快捷栏左数第三个小图标importtext data via spreadsheet interface —Load a data file(此时选择原始数据的EXCEL文件)—OK—Full matrix—保存—命名。
方法二
直接点击软件中的快捷栏左数第二个小图标matrix spreadsheet(建立UCINET本身矩阵数据表)——将原始数据粘贴过来——点击save保存数据-命名。
原因:
原始数据中,我们建立了包含文献作者和关键词的矩阵,其中,若该作者的文献中有某个关键词,那么我们就在这个作者所在行与该关键词所在列的交叉处,填上1。但是,如果我们要研究的是某个领域的研究热点,那就需要建立“关键词”-“关键词”的1-mode网络。
操作步骤:
Data—Affiliations(2-mode to 1-mode)—Input dataset(选择上一步后缀名为“.##h”的DL格式文件)—Column—OK。
操作步骤:
Network—centrality—Degree—Inputdataset(选择上一步保存的后缀名为“Aff.##h”的文件)—OK。
得到的结果如图:
操作步骤:
Visualize—NetDraw—file—open—ucinetdataset—Network—“.##h”—OK。
得到结果如图:
操作步骤:
Analysis—centralitymeasures—Set Node Sizes by—Degree。
得到结果如图:
通过第五步中生成的图我们可以看到,节点“应急能力”和“企业信誉”的图标面积最大,这表明,这两个节点的中心度高于其他节点,相应地,“应急能力”和“企业信誉”也是该领域论文中的热门关键词。
英文学习:
01 Social Network Analysis
Social network analysis is a method of sociological research. Sociological theory believes that society is not composed of individuals but of networks. The network contains nodes and the relationship between nodes. Analyze and explore the structure and attribute characteristics of the network including the individual attributes in the network and the overall network attributes. The analysis of individual network attributes includes: point degree centrality proximity centrality etc.; the overall network attribute analysis includes small-world effects small group research Condensed subgroups and so on [1].
02 Introduction of UCINET software
As one of the most popular social network analysis software UCINET software is commonly used for one-dimensional and two-dimensional data analysis. UCINET software can read text files KrackPlot pajek Negopy VNA and other format files. It can handle 32767 network nodes. Of course from the actual operation point of view when the number of nodes is between 5 000 and 10 000 some programs will run very slowly [2].
03 A small case
Usually the most important step before conducting social network analysis is to find the original data.
For example to study research hotspots of domestic scholars in the field of supply chain scholars can choose CNKI as a document database search with "supply chain" as a key word or title and obtain the best results through multiple screenings of documents. The most academic literature with reference value.
By establishing a matrix of literature authors and literature keywords it is a 2-mode network of authors and research questions. This is the raw data we need for social network analysis.
Below we carry out the specific operation of social network analysis. The data used in the case comes from reference [2].
The first step: Import the original data.
Method 1: Click the third small icon from the left in the shortcut bar in turn importtext data via spreadsheet interface — Load a data file (select the EXCEL file of the original data at this time) — OK — Full matrix — save — name.
Method 2: Directly click on the second small icon matrix spreadsheet (create the UCINET matrix data table) in the shortcut bar in the software-paste the original data-click save to save the data-name.
The second step: 2-mode to 1-mode.
Reason: In the original data we built a matrix containing the author and keywords of the document. If there is a keyword in the author’s document then we will be at the intersection of the author’s row and the column of the keyword. Fill in 1. However if we want to study a research hotspot in a certain field we need to establish a "keyword"-"keyword" 1-mode network.
Operation steps: Data—Affiliations (2-mode to 1-mode)—Input dataset (select the DL format file with the suffix ".##h" in the previous step)—Column—OK.
Step 3: Generate numerical centrality.
Operation steps: Network—centrality—Degree—Inputdataset (select the file with the suffix "Aff.##h" saved in the previous step)—OK.
Step 4: Draw the center degree graph.
Operation steps: Visualize—NetDraw—file—open—ucinetdataset—Network—".##h"—OK.
Step 5: Draw the final center degree graph.
Operation steps: Analysis—centralitymeasures—Set Node Sizes by—Degree.
Step 6: Analyze the results.
From the graph generated in the fifth step we can see that the icon area of the nodes "emergency capability" and "corporate reputation" is the largest which indicates that the centrality of these two nodes is higher than that of other nodes and accordingly "emergency capability" And "corporate reputation" are also popular keywords in papers in this field.
由于篇幅有限,今天的分享就到这里啦!
感兴趣的同学可以留言与小编交流,
咱们下周见!
参考资料:
[1]https://baike.so.com/doc/5119081-5348155.html.
[2]周勤师姐的学习文档.
英文翻译:Google翻译。
本文由LearningYard学苑整理并发出,如有侵权,请联系删除!