透過您的圖書館登入
IP:18.188.142.146
  • 學位論文

Web搜尋引擎內容偏見之研究

A Study on Content Bias of Web Search Engines

指導教授 : 楊正仁
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


隨著網路上資訊的快速成長,整個網路空間中的網頁數量變得十分龐大,搜尋引擎已經變成溝通使用者與網際網路之間最重要的媒介。然而,當使用者在使用搜尋引擎查詢資料時,皆有可能會受到搜尋引擎所存在的偏見與偏好的影響而不自覺。過去的研究中,針對搜尋引擎偏見量測所做的研究並不多。在2002年時,學者Mowshowitz與Kawaguchi的研究指出搜尋引擎存在有索引偏見與內容偏見,並且當中針對索引偏見進行量測以代表搜尋引擎的偏見現象。 然而,透過觀察搜尋引擎的內容,我們認為索引偏見的評估只呈現出部分的偏見情形。例如,當兩個搜尋引擎呈現相同索引偏見值時,他們所搜尋到的網頁內容卻不盡相同,甚至可能相差很多。 此外,搜尋引擎可能對某些特定的網頁內容過度偏重,因而導致雖然搜尋引擎存有內容偏見,但是卻無法在索引偏見的評估結果中呈現出來。 在本論文中,我們提出一個新的偏見評估機制,並且針對網路上十個知名的搜尋引擎進行評估。他們分別為About, AltaVista, Excite, Google, Goto (Overture), Inktomi, Lycos (Fast), MSN, Teoma (Ask Jeeves), 與Yahoo。透過內容相似度的評估方式,我們量測出搜尋引擎的內容偏見。配合索引偏見與內容偏見的結果分析,我們提出不同的偏見量測結果,並且依據他們的結果屬性分成四個類別以提供使用者做為進一步的參考。

並列摘要


To date, Web search engines have become an essential gateway to the vast Web space. Accompanied with the use of different technologies and strategies in Web search engines, there exists the bias issue on them. However, very few research studies focus on the issue of bias assessment. In 2002, Mowshowitz and Kawaguchi pointed out that bias is exhibited in the form of indexical bias rather than content bias. However, we argue that assessing indexical bias provides only a partial observation to bias in Web search engines. For example, when two search engines get the same result of indexical bias, their content may not be the same and even greatly differ to each other. Besides, although search engines may put undue emphasis on certain page content and result in the existence of content bias, content bias cannot be shown in the assessment of indexical bias. In our study, we first propose a mechanism to assess content bias in Web search engines. Then the experimental results of assessing both content bias and indexical bias of ten popular search engines are presented. The search engines are classified into four categories on the basis of both assessment results. Such classification is an effective reference for Internet users to choose a suitable Web search engine.

參考文獻


[5] Frans Birrer. “Understanding Values and Biases in I.T.”. ACM SIGCAS Computers
[6] Sergey Brin and Lawrence Page. “The Anatomy of Large-Scale Hypertextual Web
[7] Batya Friedman, Eric Brok, Susan King Roth, and John Thomas. “Minimizing Bias
[8] Batya Friedman and Helen Nissenbaum. “Bias in Computer Systems”. ACM Transaction
Search Engines”. In SIGIR Forum, Fall 2002.

延伸閱讀