Software Fault Prediction Based on Interval Type-2 Intuitionistic Fuzzy Logic System

Abstract

With the continuous expansion and innovations in modern software development, the rate at which defects are present in software is directly proportional to how sophisticated and complex the software tends to be. Software fault prediction therefore continues to remain an important area of research in software engineering especially as new modeling algorithms are still emerging. In spite of the fact that the potential implementations of fuzzy set theory in software fault prediction have been explored in the past, to the best knowledge of the authors, it has not yet examined how interval type-2 intuitionistic fuzzy sets with membership and non-membership degrees could be used with parameter optimization in this domain. Therefore, this work aims to adopt an interval type-2 intuitionistic fuzzy logic system to predict fault in the requirement phase of Software Development Life Cycle. Intuitionistic fuzzy logic system deals with uncertainty using separate degrees of membership and non-membership of an element to a set as well as hesitation index, therefore becoming more appropriate and flexible tool to deal with imprecision and vagueness in data. Experimental analyses show that the obtained software prediction outputs using interval type-2 intuitionistic fuzzy logic system are very close to the actual outputs which confirm that the proposed approach is a more realistic alternative for software fault prediction.