An Investigation on the Use of Machine Learned Models for Estimating Software Correctability

by

M. de Almeida, H. Lounis, and W. L. Melo

In Journal on Sofware Engineering and Knowledge Engineering. 1999. To appear

Abstract:

 In this paper we present the results of an empirical study in which we have investigated Machine Learning  
(ML) algorithms with regard to their capabilities to accurately assess the correctability of faulty software 
components. Three different families algorithms have been analyzed: Divide and conquer (top down 
induction decision tree), covering, and inductive logic programming (ILP). We have used (1) fault data 
collected on corrective maintenance activities for the Generalized Support Software reuse asset library 
located at the Flight Dynamics Division of NASA's GSFC and (2) product measures extracted directly 
from the faulty components of this library. In our data set, the software quality models generated by both 
C4.5-rules (a divide and conquer algorithm) and FOIL (an inductive logic programming one ) presented 
the best results from the point of view of model accuracy.

Paper in Acrobat Reader Format (PDF)