隨著網際網路的蓬勃發展1使用者利用搜尋引擎查詢網路資源已相當普遶， 因此藉由分析Web搜尋引擊的各種電腦交易記錄〔TransactionLogs)，對瞭解 網路使用者的資訊行為應有莫大助益。本文即是以搜尋引擎的檢索詞彙記錄分 析（SearchTeimLogsAnalysis)為基礎，嘗試觀察台灣地區網路使用者的一些 檢索行為特性，包括如檢索詞彙的特性與使用者的檢索需求等，同時探討檢索 詞彙分類時的一些問題並提出心得。

With search services over the Web as major sources of infonnationj there is an increasing demand to learn more about what people need and how they search on the net. Through analysis of transaction logs? certain users? search behaviors could be observed. This paper presents the analysis of search term logs from two real-world search engines in Taiwan. Some preliminary results have been obtained, including characteristics of users and their searching needs. Several problems with building taxonomic schemes and classifying search terms are also discussed. The findings ane valuable for understanding network information retrieval research and library & information service applications.