Summarizing Observational Client-side Data to Reveal Web Usage Patterns

摘要

Client-side event logs may reveal patterns of usage of Web pages. Nevertheless, extracting useful and novel information from this voluminous data set is a challenge for evaluation tools, since a few minutes simple task may result in a sequence of hundreds of events. This work contributes with a technique to process these logs and build a Web page's usage graph summarizing statistical information of the Web page usage concerning one or more sessions. This graph reveals patterns of real usage data, which Human-Computer Interaction specialists may find useful for inspecting accessibility and usability issues. Moreover, Web usage miners can reuse the usage graph to apply other techniques to discover other patterns or rules.