Mu-hsuan Huang

本文以Dialog系統之書目資料庫為檢索系統，檢索者的來源則為本身具有資訊需求的國立臺灣大學教職員生·本文擬對線上資訊尋求行為進行系統化的分析，其目的在探討檢索者線上檢索時之錯誤行為及其發生位置，分析零筆資料的原因，了解時間經驗對錯誤行為的影響，並比較中美終端使用者線上資訊尋求行為之異同·

The purpose of this article is to analyze the searching behavior of end users.　　　　The subjects are real end users from National Taiwan University. This article　　　uses the DIALOG command language as its searching tool. It attempts to 　　　　investigate reasons for errors and location of errors. In addition it tries to　　　　identify reasons for zero posting and compare the behavior difference 　　　between Chinese and American end users.

本文探討光碟資料庫終端使用者之檢索問題對資訊尋求和資訊檢索之影響。本研究以國立臺灣大學學生為研究對象，以PsycLIT光碟資料庫為檢索系統，計有52位檢索者，總共進行58次檢索。本研究系統化地蒐集終端使用者之檢索背景、檢索問題和檢索結果，對於檢索問題進行性質上的分析，並進而探討檢索問題對檢索結果的影響。

This article analyzes the information seeking and information retrieval behavior of undergraduate and graduate students at National Taiwan University when using PsycLIT database. 52 subjects were studied through the semester, carrying out 58 searches. This study investigates information seeking and information retrieval by analyzing the use and influence of questions.

本研究之目的在分析資訊需求者與計畫檢索者對於相同檢索問題在分類上的差異。本研究由10位資訊需求者提出檢索問題並進行檢索，再由31位計畫檢索者針對此10個問題進行59次檢索。本文主要在比較此10位資訊需求者與31位計畫檢索者對同一檢索問題在分類上的差異，並探討此差異對檢索行為所造成之影響。本研究對檢索問題分類上的差異係從檢索問題之複雜性、專指性、前提假定三構面進行分析。本研究發現，相較於資訊需求者，計畫檢索者認為同一檢索問題所認知的複雜性較高，專指性亦較高，並傾於直接將前提假定用來進行檢索。

The project is designed to analyze the difference of question classifications between two groups of people: the end-users who have real information needs and the prohect searchers. Ten end-users address their search questions and proceed to do the search, while 31-project searchers conduct 59 times of searching on the ten questions posed by the end-users. The main purpose of this paper is to compare the classification differences of the search questions in terms of their complexity, specificity, and presupposition. The result of this study reveals that the project searchers not only perceive the questions in a more complex and specific way but also tend to proceed the search directly by using presuppositions of questions.

本研究目的在探討國小高年級學生之網路資訊尋求行為，以臺北市河堤國小與市北師實小32名學童為研究對象，研究方法主要為深度訪談與關鍵事件回憶法。研究結果發現：學童最主要的網路搜尋動機為學校作業；其檢索問題多為與學校作業相關者；學童主要的資訊資源管道為網路與父母；且最常利用的搜尋引擎奇摩；在資訊檢索過程中，學童較常利用一次檢索中的直接輸入檢索問題與擷取關鍵詞輸入等方式來查找資料；另一次檢索中的直接輸入網址與點選搜尋引擎之分類目錄的檢索成功率較高；根據訪談結果，方便好用為學童提出網路搜尋資訊之最大優點，找不到資料為最大缺點。

This article investigates Internet information seeking behavior of elementary school students in their senior years from “Heti Elementary School” and “Taipei Municipal Teachers College Experimental Elementary School. ”Subjects are 32 senior grades students selected from the above two schools. “In-depth interview” and “critical incident technique” are used as the primary research methods. From the result of this study, we can infer that “school assignments” comprises the primary motive of elementary school students in their senior years. It is for this reason that most questions are asked by these pupils in relation to “school assignments.” Pupils frequently used “the Internet” and “parents” as their primary information resource channels. ”Kimo” was the search engine with the highest usage. During the information search process, pupils tended to perform “a single search” by “directly entering the search question” and “extracting the keyword and entering it.” It is the “directly entering the website address” and “selecting from the search engine’s list of contents” both of which involved using “a single search” that get the highest success rate. In addition, most of pupils believed that “convenient to use “ was the greatest advantage in using the Internet to search for information, while “could not find data” was the most commonly observed isadvantage.

本研究之目的係利用專利引用網路，探討我國資訊電子業研發領域之變遷，以及公司與公司間之關聯性。研究對象計有59家公司，根據檢索結果，59家公司於1992年至2000年間所擁有之美國專利數，總計有5,513篇，共包含引用專利數32,945篇。本研究採用文獻分析法、資料庫檢索及書目耦合分析等研究方法，以三年為一時段，對研究對象之引用專利進行書目耦合分析，再利用集群分析及多維尺度法繪製專利引用網路。研究結果顯示我國半導體業公司之關聯性最強；資訊電子業之子集合最明顯；且公司間之關係逐漸由分散走向集中，產業間之界線已愈來愈模糊，甚至有重疊的現象。

The purpose of this paper is to discuss the transition of Taiwan's information and electronic companies, and explore the relationship between companies and industries through plotting out patent citation maps. The fifty-nine companies studied in this paper, obtained 5,513 patents from U.S. between 1992 and 2000, and they cited 32,945 patents during these years. The research methods of this study include literature review, database retrieval, and bibliographic coupling. Resear-chers divided the period into three, and used cluster analysis and multidimensional scaling to plot out patent citation maps. The result of this study shows that a higher similarity among companies in semiconductor sector, and the relationship between companies is getting closer. In addition, the distinction between industries grows more and more ambivalent, even overlapping in some cases.