RECONSTRUCTION OF THE WEB APPLICATION HYPERTEXT MODEL USING WEB LOGS

摘要

An approach for automatic reconstruction of a part of a web application description, namely the hypertext model, using web logs is presented in the paper. The approach includes an advanced analysis of a large amount of the web logs data and the construction of the Application Transition Graph (ATG) as the formal description of hypertext model. Although the proposed reconstruction is limited on the navigational structure of the web application, the defined process does not depend on any particular technology and can be performed in an automatic way. The obtained results are (1) a complete model of those parts of a web application that are actually in use and (2) a very good starting point for obtaining a complete model including the unused parts as well. In either case, the approach significantly reduces the complexity and costs of reverse engineering. As a case study the approach is applied on the web student information system eStudent from the University of Ljubljana, Slovenia.