Software refactoring at the function level using new Adaptive K-Nearest Neighbor algorithm

Software architecture decomposition using adaptive K-nearest neighbor algorithm

“…A requirement attribute can be: (i) items, typically objects, explicitly specified in the requirements; (ii) user-defined attributes (e.g., data structures defined by the designer); or (iii) system attributes that relate to the operating environments or other systems. The techniques presented in [1,2] do not give this flexibility in first two steps as it is handles specific types of software components (functions or [5] packages) and specific types of the attributes for those components (features of functions or packages). For example, the refactoring approaches presented in [1,2] cannot be applied on systems written on structural programming languages while the approaches presented in [5,6] is independent from the type of programming language.…”

“…The techniques presented in [1,2] do not give this flexibility in first two steps as it is handles specific types of software components (functions or [5] packages) and specific types of the attributes for those components (features of functions or packages). For example, the refactoring approaches presented in [1,2] cannot be applied on systems written on structural programming languages while the approaches presented in [5,6] is independent from the type of programming language. Another instance is that approaches in [1,2] cannot utilize the Software Requirements Specification (SRS) for a system while the one in [5] can.…”

“…For example, the refactoring approaches presented in [1,2] cannot be applied on systems written on structural programming languages while the approaches presented in [5,6] is independent from the type of programming language. Another instance is that approaches in [1,2] cannot utilize the Software Requirements Specification (SRS) for a system while the one in [5] can. The third step is to identify the relationships between requirements and attributes.…”

“…In this paper, we investigate the use of A-KNN in addition to HAC methods while applying those five steps. A-KNN has been proposed in [2] and used for software refactoring at the function level to assist software designer in splitting large software functions into smaller and more cohesive ones but it was not applied at architecture level at that study as it depends on features for functions. A-KNN has been used for software refactoring at the package level [1], however, the objective in that study is limited to balance software coupling and cohesion by redistributing classes over packages or create new packages in software systems written using objectoriented programming languages.…”

Implication of Clone Detection and Refactoring Techniques using Delayed Duplicate Detection Refactoring

“…Technical management and requirements have the less importance on humans [8]. The Adaptive K -Nearest neighbor clustering algorithm is used to perform the clustering in ill-structured software entities refactoring at function level [16]. The pull up method refactoring is used to eliminate the duplicate codes in subclasses.…”