Performance of Activation Function Based Clustering of Automated Software Coding Contracts

Abstract

The conceptualization of software coding related Contracts where programmers’ involvement being minimalist is very decisive to pragmatically give rise to contracts optimized enough mitigating the issues in adhering to programs specifications, and setting the stage for extensive and augmented testing of various real world software. The reference framework introduced incipiently draws up specifications of dependency in terms of behavior in conjunction with structural aspects in the form of restraints or conditions structured over a Decision tree eventually being materialized as Automated Programming Contracts. The nuances of refining the extracted contracts are further endorsed to enhance them for strengthened verification proficiencies. This is realized using an improved variation of K-Means Clustering method to establish meticulous clusters of assertions/contracts comparable to one another. The clustering performance of proposed algorithm is compared with that of certain existing algorithms.  Further efforts on this track could be made to classify the contracts and CPU and Memory related operations getting better with effective capabilities to expose the software errors.