Evaluation of Website Usability Using Markov Chains and Latent Semantic Analysis

摘要

The development of information/communication technology has made it possible to access substantial amounts of data and retrieve information. However, it is often difficult to locate the desired information, and it becomes necessary to spend considerable time determining how to access specific available data. This paper describes a method to quantitatively evaluate the usability of large-scale information-oriented websites and the effects of improvements made to the site design. This is achieved by utilizing the Cognitive Walkthrough for the Web and website modeling using Markov chains. We further demonstrate that we can greatly improve usability through simple modificacion of the link structure by applying our approach to an actual informational database website with over 40,000 records.