An enhanced cybersecurity testing based on CHSPNN and DG in software testing for banking or finance applications

Cyber testing is very essential for banking or finance applications to test the software for the identification of cyber-attacks in the software codes. None of the existing work in financial software has covered layers such as network breaches, application layer attacks, and social engineering and insider threats for cyber-security testing. Therefore, the paper presents cyber-security testing in software testing, covering all these layers. Initially, the input software code is converted into a graphical representation using AST. From this, vectorization and patching is carried out by B-SMART and SRSKFC, followed by feature extraction. After the patching process, graph kernel analysis is done, and with the extracted features, both are given to feature selection performed by GSCLO to select the optimal features. Then, behavior analysis takes place, which consists of pre-processing the data, and from this data, features are extracted. Now, extracted features and selected optimal features are given to CHSPNN classifier. Finally, from the obtained class the attacked surface in the code is mapped using DG. Hence, the proposed model is classified and the attacks are mapped very effectively, with an accuracy of 97.97%.