Performance Evaluation of Data Mining Techniques to Enhance the Reusability of Object-Oriented (O-O) Systems

Abstract

The software industry is evolving at a rapid pace, making it necessary to optimize efforts and accelerate the software development process. Software can be reused to achieve quality and productivity goals. Reusability is a crucial measure that can be used to increase the overall level of software quality in less time and effort. To better understand the necessity of enhancing the software reusability of Object-Oriented (O-O) systems, this study employed a semi-automated approach to measure the values of class-level software metrics on an input dataset collected from the MAVEN repository. This paper explored several previous studies, data strategies, and tools to predict reusability in O-O software systems. This study compares various data mining techniques to identify the most suitable approach for enhancing the reusability of O-O software systems. The analysis was based on performance parameters such as precision, MSE, and accuracy rates. Due to its higher precision and lower MSE, the SOM technique is considered one of the top data mining approaches to increase the reusability of O-O software systems. However, the results show that the different levels of reusability in O-O software systems are not adequately addressed in current solutions.