Makine öğrenmesi ile mobil uygulama sınıflandırılması ve otomatik keşif testi (Mobile application classification using machine learning and automated exploratory testing)

This is the latest version of this item.

Official URL: http://ceur-ws.org/Vol-1980/YTM_2017_paper_12.pdf

The knowledge of the business domain of a Software-Under-Test (SUT) is crucial for testing. Therefore identification of business domain and the underlying business processes is the basis for automated testing. Test cases and test input set can be automatically generated depending on the domain and process information. In this research, we apply machine learning techniques to determine the similarity of applications. Applications in the same domain should be highly similar and we can say that, same business processes are implemented in the applications of a business domain. Our hypothesis argues that assuming we can create a generalized Finite State Machine (FSM) model of a business domain, the states and transitions of the FSM could be matched to the business processes of a business domain. Previously created test cases and test input could be used for testing an application that is coherent with the states and transitions of the formal model. In this research we coin two novel terms,Model Dressing and Automated Exploratory Testing. Model dressing is matching an application to the generalized model of a business domain. Automated exploratory testing is using the previously gathered business domain knowledge to test new applications and gradually merging outcome to the previous know-how to improve testing process.