相關排序於資訊檢索之發展與探討

Development and Issues Concerning the Relevance Ranking for Information Retrieval

張郁蔚(Yu-Wei Chang)

《大學圖書館》 8卷2期 (2004/09) Pp. 94-123

摘要

面對網路資源檢索結果往往過多的問題，如何將當中最相關的資訊優先排序在最前面，讓使用者優先看到，以節省其瀏覽所有檢索結果的時間，是目前系統開發者選擇相關排序（Relevance ranking）作為檢索結果排序的原因。目前大部分的資訊檢索系統係以相似性而非相關性作為檢索結果排序基礎，但使用者係判斷檢索結果相關與否之唯一裁判者，為此，相關排序的發展也開始將使用者的相關判斷及資訊使用行為作為改善相關排序效能之努力方向，希望能針對使用者的不同需求提供更個人化的資訊服務。

關鍵字

相關 ； 排序 ； 相關排序

並列摘要

The retrieval results of a web search are usually very large. Therefore, the information retrieval system usually uses the relevance ranking to ranking retrieval results, provideing users a saving-time way to look through the searching results and obtain the most relevant documents. Currently, most information retrieval systems use similarity instead of relevance to rank the searching results. However, the end-user is the one to determine the relevance of retrieval results, so many researchers consider the user's relevance requirement and information behavior are the way to improve the effectives of relevance ranking. The objective of relevance ranking is to provide personalized information service for users.

並列關鍵字

Relevance ； Ranking ； Relevance ranking

參考文獻

黃慕萱(2001)。圖書資訊學概論。

Relevancy Ranking

Andreas Paepcke,Hector Garcia-Molina,Gerard Rodriguez-Mula,Junghoo Cho(2000).Beyond Document Similarity: Understanding Value-Based Search and Browsing Technologies.ACM SIGMOD Record.29(1)

Behnak Yaltaghian,Mark Chignell(2002).Re-ranking Search Using Network Analysis: a Case Study with Google.in Proceedings of the 2002 Conference of the Centre for Advanced Studies on Collaborative Research.2

Caroline M. Eastman,Bernard J. Jansen(2003).Coverage, Relevance, and Ranking: The Impact of Query Operators on Web Search Engme Rusults.ACM Transactions on Information System.21(4)