The statistical analysis of source-code to determine the refactoring opportunities factor (ROF) using a machine learning algorithm

In this research, we have proposed refactoring as a measured object with the help of a measurement scale. We have proposed a case study in which we have studied three different projects, obtained from a company for which an ordinal scale is prepared. The UML diagrams are drawn from which the values of different source-code metrics, those are helpful to determine the quality of the code, are calculated. A refactoring opportunities factor (ROF) has been introduced which determine the correct perspective of refactoring. Each UML diagram is assigned a ROF based on the values of source-code metrics. A machine learning algorithm is developed, based on the Naive Bayes Algorithm, which takes dataset prepared by studying 3 projects, as an input and determines which of these has good and bad opportunities for refactoring. The accuracy (precision and recall) of the machine learning classifier validates the refactoring opportunities factor.