With the ever increasing number of software applications and the critical risks which arise due to the low quality software, the utter importance of high quality software development gets to be more important than ever. The model that we have defined and illustrated here tries to provide an explicit process for ultimately binding quality-carrying properties into the desired software. These properties simply imply particular quality attributes in turn. In our research work, we proposed a quantitative Clone Detection model with respect to the Component Based Development (CBD) methodology using SOFMs. We have used the application of C. K. matrices so as to find out the cloning structure from each and every class of various types of design patterns (components). While we add up the matrices value for each and every component, the number of classes is also calculated at exact run time and concated with the final matrices value. This in total acts as an input for our neural network. Then neural network is used to train the self organizing map (SOM) and then the neural networks are used to find out in which case, maximum performance can be achieved. By using both the examples of design patterns and unsupervised neural network, we have proposed a dynamic model that provides far better performance for software component quality model.
Clone Detection, Clone clustering, Software Matrices, Cohesion and Coupling, SOM, SOFM
Himanshu Chaudhary, Ramesh Belwal, An Algorithmic Approach for Clone Clustering using Software Matrices and SOFM's, HCTL Open International Journal of Technology Innovations and Research (IJTIR), Volume 12, December 2014, e-ISSN: 2321-1814, ISBN:978-1-62951-791-9.