LAUSR

Software requirements with its functional and non-functional methods are the first important phase in producing a software system with free errors. The functional requirements are the visual actions that may easily evaluated from the developer and from the user, but non-functional requirements are not visual and need a lot of efforts to be evaluated. One of the main important non-functional requirements is security, which focuses on generating secure systems from strangers. Evaluating the security of the system in earlier steps will help to reduce the efforts of reveals critical system threats. Security threats found because of leaking of security scenarios in requirement phase. In this paper, we purpose an intelligent model to extract and evaluate security features from scenarios based on set of security system goals and a set of security requirements saved on rich story scenarios dataset (RSSD). This model will used a support vector machine (SVM) classifier to classify the security requirement based on RSS dataset.  The using of SVM will enhance the overall process of evaluating the security requirements. The results show a significant enhancement in security improvements.