網際網路的應用掀起一陣標籤(Tags)熱潮,世界各國的網站無論是影音串流(Video Streaming)或相簿(Album)分享等資源分享網站,皆將標籤的概念加入並強調資源共享的機制,透過網路使用者自發性的提供標籤或關鍵字(Keywords)來達到分類及共享之目的。所謂的「標籤行為」或「下標籤」(Tagging)是指網路使用者賦予網路資源一組標籤或關鍵字的行為。根據過去研究得知,網路使用者在下標籤的主要動機可分為個人因素和社群因素,個人因素多以進行個人資訊管理、檢索需要為主要動機;社群因素多以分享、吸引注意力為主。 標籤在網路上被網路使用者所標記的數量隨著時間的增長而不斷增加,因而許多網站便將標籤匯聚成標籤雲(Tag cloud),故許多學者開始針對這些資訊進行研究。由過去的研究可看出大多數關於標籤的研究可分三大類:(1)探討標籤的使用模式研究;(2)標籤的語意處理研究;(3)標籤的類別研究 [4],不過上述研究鮮少探討標籤隨時間變化之分析研究。標籤隨時間的變化可反映出網路使用者的興趣或偏好改變,若加以觀察可發現此現象之變化情形。 有鑑於此,本研究期許能運用標籤隨時間的變化,協助分析網路通用之熱門標籤內容及其隱含之潛在內容,期許能幫助企業找出對網站行銷有幫助之訊息。因此本研究提出一個標籤變化演算法(Algorithm of Tag Variation, ATV)來分析標籤的變化情形,將熱門網站的網址與國際知名Delicious網站的標籤資料做彙整與分析,由標籤的變化趨勢來找出標籤的淡入(Fade in)、淡出(Fade out)與持平(Stay still)的狀態。由本研究之分析與探究,找出下列網路標籤的現象:(1) 標籤的變動週期約為3~5天;(2) 每個網址的Top30標籤多以Top1~5的標籤被標記次數居多;(3) 標籤應用於網址上多以呈現其網站的代表性功能為主。另外,由本研究所提的ATV方法具有下列特色:(1)將標籤隨時間改變之變化趨勢做分析,將不適用的標籤或熱門標籤做區隔(僅採TOP 5的標籤),如此可提高標籤的參考價值與實用性;(2) 區分標籤的狀態後,可將不適用的標籤從資料中排除,降低標籤資料的不準確性;(3)此外,也可從標籤的變動趨勢找出對企業或網站經營者有幫助的資訊,挖掘標籤對企業行銷影響的潛在性。 根據本研究的實驗結果,本研究發現標籤隨時間變化的特有現象,並給予分析結果初步探討,也提供一標籤分析流程幫助企業或網站經營者可掌握網路使用者之標籤變化,了解使用者的認知與偏好。
Tags are textual description defined by users and are prevalent in most web 2.0 website. With the growth of tag applications on internet, more and more websites, like the video streaming (Youtube.com) or album sharing (Filckr.com), are all taking the concept of tagging in consideration. The usage of tagging is treated as the classification and sharing for these tagged resources. The“tagging behavior” or “tagging” means the process of assigning personal descriptive keyword or phrase to a web resource. According to the passed studies, there are two main motivation of tagging, the first is the personal factor and the second is the social factor. The personal factor is like personal information management or information index. The social factor is treated as sharing or attracting other people’s attention. The amount of tags are usually increased over time, and many websites collect these tags to form the tag cloud. As the tagging become increasingly popular, there are also a growing amount of study on this area. These researches are focus on three subjects: (1) the usage of tag; (2) the semantics problem of tag, and (3) the types of tag. However there are few researches focus on the variation of tag within a period of time. The variation of tag can reflect the change of preference or interests of users. If we can observe the tag’s variation, then the change and the trend of user’s interest can be found. This research uses the variation of tag to analyze those popular tags on the internet, and finds out the potential content for enterprises on the marketing of internet. This research proposes an Algorithm of Tag Variation (ATV) for analyzing. The popular website URL is used as the research domain and the tag data is collected from Delicious.com for experiment. The tag status like Fade In (FI), Fade Out (FO), and Stay Still (SS) are defined in this research. According to the experimental results, we discover some phenomena, (1) the tags variation cycle is about 3 to 5 days, (2) the top 1 to top 5 tags are those tags that have more variation of count than others, and (3) tags can reflect the potential meaning or function of a website.