A Tool-Supported Process for Reliable Classification of Web Pages

Reliable classification of Web Application User Interfaces for the aim of extracting specific data for each class of interfaces is a fundamental task in migration, testing and reverse engineering processes involving existing Web Applications. A feasible and reliable classification approach is the one that exploits combinations of Web pages structural features for discriminating the page equivalence class. This paper presents a technique based on an iterative process that allows classification rules composed of Web pages structural features to be deduced in dynamically generated web pages. The process is supported by a tool that partially automates the process steps. In order to assess the process feasibility and cost effectiveness, a case study addressing the problem of generating classification rules for a real Web application has been carried out.