Artificial Intelligence Model for Software Reusability Prediction System

Abstract

The most significant invention made in recent years to serve various applications is software. Developing a faultless software system requires the software system design to be resilient. To make the software design more efficient, it is essential to assess the reusability of the components used. This paper proposes a software reusability prediction model named Flexible Random Fit (FRF) based on aging resilience for a Service Net (SN) software system. The reusability prediction model is developed based on a multilevel optimization technique based on software characteristics such as cohesion, coupling, and complexity. Metrics are obtained from the SN software system, which is then subjected to min-max normalization to avoid any saturation during the learning process. The feature extraction process is made more feasible by enriching the data quality via outlier detection. The reusability of the classes is estimated based on a tool called Soft Audit. Software reusability can be predicted more effectively based on the proposed FRF-ANN (Flexible Random Fit -Artificial Neural Network) algorithm. Performance evaluation shows that the proposed algorithm outperforms all the other techniques, thus ensuring the optimization of software reusability based on aging resilient. The model is then tested using constraint-based testing techniques to make sure that it is perfect at optimizing and making predictions.