Natural Language Processing and Deep Learning Towards Security Requirements Classification


Security Requirements classification is an important area to the Software Engineering community in order to build software that is secure, robust and able to withstand attacks. This classification facilitates proper analysis of security requirements so that adequate security mechanisms are incorporated in the development process. Machine Learning techniques have been used in Security Requirements classification to aid in the process that lead to ensuring that correct security mechanisms are designed corresponding to the Security Requirements classifications made to eliminate the risk of security being incorporated in the late stages of development. However, these Machine Learning techniques have been found to have problems including, handcrafting of features, overfitting and failure to perform well with high dimensional data. In this paper we explore Natural Language Processing and Deep Learning to determine if this can be applied to Security Requirements classification.



Software And Hardware

• Hardware: Processor: i3 ,i5 RAM: 4GB Hard disk: 16 GB • Software: operating System : Windws2000/XP/7/8/10 Anaconda,jupyter,spyder,flask Frontend :-python Backend:- MYSQL