ADABOOST FUZZY SVM BASED SOFTWARE DEFECT PREDICTION

Abstract

Software Defect Prediction (SDP) is a critical task in the software development process that forecasts which modules are more prone to errors and faults before the testing phase begins. This paper proposes a SMOTE algorithm to handle class balanced dataset issue. Adaboost Priority based Fuzzy SVM classification technique is proposed for classifying the balanced data sets. This proposed method reduces error rate compared with other existing machine learning and Priority based Fuzzy SVM methods. Experimental results showed that proposed scheme yields better accuracy than the existing techniques.