Abstract

In previous work, we have shown that false positive detection of widgets can be reduced by searching for the Web page’s Document Object Model (DOM) elements that the widget monitors and updates. Using the profile of the JavaScript code’s structure from a Web page, the link between the widget’s set of JavaScript code and the DOM elements that interacts with the users can be established. The profiler can also be used as a platform to include tell-sign’s detection for widget identification research to be conducted. In this report, the architecture of the Widget Identification System (WIS), the process flow of the profiler, its limitations, the data structure that stores the profile of the analysed code and the evaluation results of the profiling approach are presented. Due to the size of the full JavaScript code from the top 10 Websites’s default pages, the techniques and assumptions made to overcome the scale of the evaluation are discussed. The profiling approach achieved a 100% detection accuracy from the evaluation, thus demonstrating the reliability of the platform.