科學文獻之合著及熱門主題網絡分析：以機器人文獻為例

The Co-authorship and Hot Topic Network Analysis of Robots Literatures

許雅珠(Ya-Chu Hsu)

指導教授 ： 黃鴻珠 ； 黃元鶴

淡江大學/文學院/資訊與圖書館學系碩士班/碩士(2011年)

https://doi.org/10.6846/TKU.2011.00748

摘要

本研究主要探討機器人文獻之合著情況及高被引文獻之熱門研究主題，以科學引文索引擴充版資料庫中14種機器人學核心期刊文獻為研究資料來源，樣本資料共計9,571篇研究論文，分布於1983-2009年間。首先，應用書目計量分析，探討機器人文獻之合著整體概況、跨國合著、跨機構合著與作者合著情形，以及進行高被引文獻分析與關鍵字分析。其次，以社會網絡分析中之三種中心性檢測最具資源控制能力之國家、機構及作者，並繪製合著與高被引文獻熱門研究主題之網絡圖。最後，將關鍵字進行集群分析與多維尺度分析。

本研究分析結果顯示，機器人學每年研究論文篇數由96篇成長至746篇，整體之合著率為82.71%，以2至4位作者為最主要的合著型態。在國家層次方面，整體跨國合著率為11.98%，美國是跨國合著篇數最高且最具資源控制能力之國家。在機構層次方面，整體跨機構合著率為29.24%，美國卡內基美隆大學是跨機構合著篇數最高之機構，而美國麻省理工學院為最具有資源控制能力之機構。在作者層次方面，Vukobratovic是合著篇數最高之作者，而Fukuda與Cho則為較具資源控制能力之作者。高被引文獻之關鍵字分析結果呈現六大研究主題，包括無固定路線之機器人、智慧型機器人、視覺機器人、固定路線或固定點之移動機器人、時域機器人、夾爪機器人。透過上述之分析結果，可呈現機器人學整體研究概況，以提供產官學相關單位參考。

關鍵字

合著 ； 共字分析 ； 書目計量 ； 社會網絡分析

並列摘要

The purpose of this study is to explore the co-authorship and the hot topic from the highly cited articles of robots literature. The research targets are 14 journals of robotics from Science Citation Index Expanded (SCIE) database. A total of 9,571 articles were collected from 1983 to 2009. First of all, the bibliometric approach was applied to explore the co-authorship with individual, institutional and international levels, as well as highly cited articles and keyword analysis of robots literature. Secondly, the social network analysis was applied with three measures of centrality which are degree centrality, closeness centrality, and betweenness centrality in individuals, institutions and countries levels, as well as the collaboration network and keywords co-occurrence network. Finally, the cluster analysis and multi-dimension analysis were conducted for the keywords of those highly cited robots literature.

The conclusions are as follows. The amount of articles from robots literature increased gradually from 96 to 746 each year. The percentage of co-authored articles is 82.71% and the most frequent numbers of authors per article are ranging from 2 to 4. At the national level, the multinational collaboration is 11.98% of the co-authored articles. The U.S. owns the largest amount of multinational collaboration articles and has the highest value of three kinds of centrality measures, which means that the U.S. plays the key role in controlling resources. At the institutional level, the cross-institutional collaboration is 29.24% of the co-authored articles. Carnegie Mellon University owns the largest amount of cross-institutional collaboration articles; however, Massachusetts Institute of Technology plays the key role in controlling resources. At the individual level, Vukobratovic has the largest amount of co-authored articles; however, Fukuda and Cho play the key role in controlling resources. The result of keywords analysis of highly cited articles shows 6 major research categories including flexible-path of robot, intelligent robot, vision robot, fixed path or point of mobile robot, time-domain robot, and end-effect robot. The above analysis presents the summary of robotics, which might be helpful for policy making in promoting robotics.

並列關鍵字

co-authorship ； co-word analysis ； bibliometric ； social network analysis

參考文獻

林秀惠（2007）。從期刊文獻的引用分析學科領域的內涵與影響：以資訊科學為例。未出版之碩士論文，淡江大學資訊與圖書館學研究所，臺北縣。

林利真（2007）。我國電機電子領域期刊文章合著之研究。未出版之碩士論文，國立臺灣大學圖書資訊學研究所，臺北市。

潘同泰（2010）。機器人發展史與原理。華岡工程學報，25，1-6。

傅雅秀（2002）。從生命科學期刊論文作者數探討科學合作。圖書資訊學刊，17，71-80。

陳美惠、李來錫、王相華、蘇倍以（2008）。以社會網絡分析法進行社區生態旅遊發展之研究─以社頂部落的共識凝聚為例。中華林學季刊，41(2)，221-236。