An efficient and secure feature location approach in source code using Jacobian matrix-based clustering

Abstract

In software, a feature is a functionality, which has the chief role in identifying the primary location of the source code (SC), and it is stated by means of requirements as well as accessibility to developers and also users. Prevailing feature location (FL) methods have issues in security effectiveness. This paper proposed an effective and safe FL approach in SC utilizing the Jacobian matrix-based clustering (JMC) to trounce these issues. Originally, the requirement-based approach removes the repeated test cases (TC). Subsequently, the weight-based salp swarms algorithm (Ψ-SSA) selects the significant TC attributes. Followed by which, the associations rule mining (ARM) process is implemented. The ARM encompasses the itemset, support, frequent itemset, closed frequent itemset, confidence, in addition to divergence. Subsequently, the affinity is computed by means of the blend of confidence with divergence. In this affinity computation, the MALO optimizes the weight value. Then, the exponent-based elliptic curves cryptographic encrypts the affinity value. Subsequently, the score value is computed for the encrypted affinity value centered on the entropy computation. Finally, the JMC locates the feature centered on the score value. Experimental assessment exhibits that the proposed system’s performance is better than that of the prevailing research methodologies.