現在網路使用者利用搜尋引擎進行關鍵字搜尋的方式越來越普遍,從實務界的應用可以得知關鍵字搜尋的重要性也越來越重要。但是鮮少有研究探討人們如何開始搜尋、以及使用者進行搜尋的動機和搜尋所使用之關鍵字為何。因此本研究的目的為探討關鍵字搜尋趨勢變化情形,以及搜尋動機、人口變數與關鍵字搜尋行為間的關係。 本研究首先將2007年上半年最熱門的1008個關鍵字依單根檢定分成穩定型、波動型、趨勢型、混沌型等四類;其次將關鍵字依照搜尋動機分成指向型、資訊型、資源型關鍵字。再根據不同類型的圖形解釋其趨勢涵義,最後利用卜瓦松回歸模型探討使用者與關鍵字搜尋間的影響變數。研究結果發現大多數變數:搜尋長度、搜尋寬度、搜尋深度、搜尋引擎使用比例、造訪網站關聯性、網站停留時間及人口統計變數:年齡、教育程度都會對關鍵字搜尋總數產生影響。 根據上述研究結果,本研究也闡述了未來關鍵字搜尋模型在網路搜尋行為上的應用,且對未來的研究提供了一個有關資訊行為變數及人口變項對各類關鍵字分類的搜尋量的預測工具基礎。
Online keyword searching becomes more popular recently. However, still very few studies focus on exploring keyword searching behavior, such as how people search? Why do they search? What keywords they are using for search? The purpose of this study is to explore online users’ keyword searching and information behavior, and to find out the relationship among online users’ search motivation, characteristics, information behavior, keywords’ types and search amount. First, this study is to classify the top 1008 keywords, being searched in first-half of 2007, into four types by unit root test. These keywords are classified into four categories: steady type, fluctuate type, tendential type and chaotic type. Second, we also find out 3 different types, navigation, information and resource of keywords, identified by online users’ searching goals. Third, this study uses the poisson regression model to explore the relationships between the keywords types and online users’ characteristics and information behavior. The results show most variables have significant effects on the amount of every certain type of keywords. Finally, this study also illustrates how the proposed keywords searching behavior model can applied the web searching in the future. This study also provided a tool to forecast keyword search trend for further theoritcal and practical studies.