Classifying Model-View-Controller Software Applications Using Self-Organizing Maps

The new era of information and the needs our society require continuous change in software technology. Changes are produced very quickly systems evolving at same velocity, which implies that decision-making process architectures should be (semi-)automated to satisfy changing avoid wrong decisions. This issue is critical since suboptimal architecture design decisions may lead high cost poor quality. Therefore, systematic mechanisms help architects during required. Architectural patterns one most important features applications, but pattern can implemented different ways, leaving results When an application requires evolve, knowledge extracted from similar applications useful for driving decisions, quality implementations reproduced improve specific attributes. clustering methods especially suitable classifying implementations. In this paper, we apply a novel unsupervised technique, based on well-known artificial neural network model Self-Organizing Maps, classify Model-View-Controller (MVC) point view. Software analyzed by 24 metrics organized into categories Count/Size, Maintainability, Duplications, Complexity, Design Quality. main goal work twofold: identify establish similarity MVC without architect bias, means Maps metrics. To end, performs exploratory study conducting two analyses with dataset 87 Java characterized attributes describe technology dimension application. stated findings provide base applications.