A Web Site Measurement Based on Information Theory

• Main Authors: Chia-Hsien Tzeng, 曾嘉賢
• Other Authors: Cheng-Tzu Wang
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/24816758522675606825

碩士 === 國立臺北教育大學 === 資訊科學系碩士班 === 97 === A web site can be evaluated in three ways, user survey, usage analysis, and navigability measurement. User survey uses a sample of a web site’s user population to assess the web site’s satisfiability against a set of criteria or a checklist. This process is time consuming and the result depends on the representative sample. Usage analysis examines web site usage data such as the log data. This method doesn’t consider the hyperlink structure of a web site. Navigability measurement quantitatively assesses the global or local properties of a web site. Typical navigability measures are based mainly on the static hyperlink structure of a web site. These measures however, have two main drawbacks, one is that the effect on navigability of a hyperlink structure cannot be well characterized, and the other is that the behavior of user’s navigation is ignored. In this thesis, a novel Markov-based web site measure is proposed. This proposed measure considers not only the hyperlink structure of a web site, but also the behavior of users’ navigation. The Markov-based measure also provides suggestions to improve navigability of the web site. Experimental results show the proposed scheme can measure web site efficiently.