Machine Learning Algorithm for Assessing Reusability in Component Based Software Development

EasyChair Preprint 4142

Abstract

Software reusability has been present for several decades. Software reusability is defined as making new software from existing one. Objects that can be reused: design, code, software framework. We reviewed several approaches in this dissertation, i.e. object-oriented metrics, coupling factor, etc., by which the software's reusability increases. Therefore this thesis analysis on how to classify and reuse the program using those metrics and apply the algorithm of machine learning. In this thesis we test open source software and generate a ck metric of that source code then a machine learning algorithm will process the data using weka tool to give the result. We test coefficient of correlation, mean absolute error, root mean square error, relative absolute error and root relative square error less the program would be better from this we get 98.64 accuracy on online examination system software.

Keyphrases: CK metric., Machine Learning Algorithm, Random Forest, Reusability

@booklet{EasyChair:4142,
  author    = {Pooja Negi and Umesh Kumar Tiwari},
  title     = {Machine Learning Algorithm for Assessing Reusability in Component Based Software Development},
  howpublished = {EasyChair Preprint 4142},
  year      = {EasyChair, 2020}}