Today, users use web applications to perform essential tasks like business and e-government. Therefore, the quality and reliability of web applications must be eva‎luated and tested from various perspectives, including security and functionality. In order to eva‎luate the quality of web applications, it is crucial to extract the behavior of web applications. Extraction of web application behavior manually is time-consuming and costly. Therefore, automated methods for analyzing the behavior of web applications are essential for eva‎luating their quality. Web application state space is so large that automated methods for analyzing web behavior cannot fully cover them because of time constraints. Thus, methods for automating the extraction of web application behavior face the challenge of adequately covering the functionalities of web applications in their vast state space. This research presents a reinforcement learning-based interactive method to extract the behavior of web applications. This implies that the reinforcement learning method interacts with the web application. Through this interaction, it will gradually cover and extract a web application's behavior and functions. Several web applications have been used to eva‎luate the effectiveness of the proposed method. Based on the eva‎luation results, it appears that the proposed method extracts the behavior of web applications more effectively and covers more web application functionalities.

مريم موزراني

مازيار پالهنگ، زينب زالي