Analysing Agreement Among Different Evaluators in God Class and Feature Envy Detection

The automatic detection of Design Smells has evolved in parallel to the evolution refactoring tools. There was a huge rise research activity regarding Smell from 2010 present. However, it should be noted that adoption real software development practice is not comparable On basis assumption objectiveness operation as opposed subjectivity definition and identification makes difference, this paper, lack agreement between different evaluators when detecting empirically studied. To do so, series experiments studies were designed conducted analyse concordance persons tools, including comparison them. This work focuses on two well known Smells: God Class Feature Envy. Concordance analysis based Kappa statistic for inter-rater (particularly Kappa-Fleiss). results obtained show there no general, and, those cases where certain appears, considered fair or poor degree agreement, according Kappa-Fleiss interpretation scale. seems confirm subjective component which raters evaluate presence differently. study also raises question training experience Smells.