A Novel Approach of Software Modules Reusability with Aging Classification by Object- Oriented Metrics with Deep Learning using Random Forest Approach

Ridhi Jindal ,S.k. Mittal

The proposed approach is expected to utilize a novel transformative computerized machine learning/artificial intelligence for regression-based tests to be utilized to estimate reusability. Such improvement can prompt exact reusability design estimation, which can be powerful for designing optimal software configuration. This approach is famously known as aging-resilient software-based reusability. The proposed framework utilizes prevalent object-oriented programming metrics, for example, Kemerer and Chid amber’s metrics to inspect reusability. Here, cumulative metrics, McCabe's metrics, objectoriented metrics, coupling, and a cohesion-based reusability evaluation approach have been proposed which could be of supreme significance in the optimization of software design. This paper involves the development of primary designing and software-metrics algorithms for assessing the metrics via UML/class structures. It is possible to determine a robust and effective reusability expectation model for web-based service products utilizing Object Oriented metrics. It also discovered that the OO-CK metrics, especially complexity metrics, coupling, and cohesion-based metrics can be useful in forecasting reusability in the web-based software frameworks. These features get upgraded by the convolution neural system (CNN) and perform the mechanism of learning by a random forest technique thereby enhancing the mechanism of reusability.

Volume 12 | 06-Special Issue

Pages: 240-249

DOI: 10.5373/JARDCS/V12SP6/SP20201029