Neuro –Genetic Approach for Predicting Maintainability Using Chidamber and Kemerer Software Metrics Suite

Abstract

Accurate estimation of attributes such as effort, quality and risk is of major concern in software life cycle. Majority of the approaches available in literature for estimation are based on regression analysis and neural network techniques. In this study, Chidamber and Kemerer software metrics suite has been considered to provide requisite input data to train the artificial intelligence models. Two artificial intelligence (AI) techniques have been used for predicting maintainability viz., neural network and neuro-genetic algorithm (a hybrid approach of neural network and genetic algorithm). These techniques are applied for predicting maintainability on a case study i.e., Quality Evaluation System (QUES) and User Interface System (UIMS). The performance was evaluated based on the different performance parameters available in literature such as: Mean Absolute Relative Error (MARE), Mean absolute error (MAE), Root Mean Square Error (RMSE), and Standard Error of the Mean (SEM) etc. It is observed that the hybrid approach utilizing Neuro-GA achieved better result for predicting maintainability when compared with that of neural network.